Effects of Freeze–Thawing and Intravenous Infusion on Mesenchymal Stromal Cell Gene Expression

Abstract

Mesenchymal stromal cells (MSC) are increasingly used as an investigative therapeutic product for immune disorders and degenerative disease. Typically, MSC are isolated from human tissue, expanded in culture, and cryopreserved until usage. The safety and efficacy of MSC therapy will depend on the phenotypical and functional characteristics of MSC. The freeze-thawing procedure may change these characteristics. Furthermore, the cells encounter a microenvironment after administration that may impact their properties. It has been demonstrated that the majority of MSC localize to the lungs after intravenous infusion, making this the site to study the effects of the in vivo milieu on administered MSC. In this study, we investigated the effect of freeze-thawing and the mouse lung microenvironment on human adipose tissue-derived MSC. There were effects of freeze-thawing on the whole genome expression profile of MSC, although the effects did not exceed interdonor differences. There were no major changes in the expression of hemostatic regulators on transcriptional level, but significantly increased expression of procoagulant tissue factor on the surface of thawed adipose MSC, correlating with increased procoagulant activity of thawed cells. Exposure for 2 h to the lung microenvironment had a major effect on MSC gene expression and affected several immunological pathways. This indicates that MSC undergo functional changes shortly after infusion and this may influence the efficacy of MSC to modulate inflammatory responses. The results of this study demonstrate that MSC rapidly alter in response to the local milieu and disease-specific conditions may shape MSC after administration.

Introduction

Mesenchymal stromal cells (MSC) are used as an investigative therapy for degenerative and immune disease. On the road to development of an effective therapy, MSC are being examined in numerous in vitro, preclinical, and clinical studies. The outcomes of these studies have so far shown to be variable and the efficacy of MSC in clinical studies does not always match the expectations raised by preclinical and in vitro findings [1].

There are a few factors concerning MSC functionality that could explain some of the discrepant outcomes. First, culture-expanded MSC may not be fully compatible with human blood and trigger the instant blood-mediated inflammatory reaction (IBMIR) when administered intravenously [2]. Second, one of the recurring differences between preclinical and clinical studies is the use of MSC from continuous cultures in vitro and in most preclinical studies, as opposed to the use of cryopreserved MSC in the large majority of clinical studies [3]. Third, the phenotype and functionality of MSC may change upon encounter with the in vivo microenvironment after administration.

The IBMIR reaction toward MSC may on the one hand compromise therapeutic cell survival, but on the other hand also trigger their beneficial paracrine effects in vivo [2]. Triggering of IBMIR results in release of factors that can activate MSC, but may also promote priming of anti-inflammatory effector cell types (e.g., regulatory T cells, myeloid-derived suppressor cells, and alternatively activated macrophages) in response to opsonized MSC [4,5].

There are indications for significant differences between MSC that are thawed shortly before use and MSC that come straight from the culture flask. Frozen–thawed MSC have impaired immunomodulatory properties and demonstrate increased triggering of IBMIR compared to MSC from continuous culture [6]. Thawed MSC furthermore show elevated levels of heat shock proteins and impaired responsiveness to inflammatory conditions within the first 24 h after thawing [7]. It is therefore possible that cryopreserved MSC are less effective than MSC from continuous culture for certain purposes, whereas for other applications cryopreserved MSC may be particularly suitable. It is important for the advancement of MSC therapy that the effects of cryopreservation on cell functionality are mapped in detail so that optimally effective MSC can be used for therapy.

Working with living cells implies that the cells can change their phenotypical and functional properties in response to environmental stimuli. The in vivo milieu that cells encounter upon administration may influence cellular function. However, insufficient knowledge of the homing habits of administered MSC, limited cell survival, and the complexity of the in vivo environment make it difficult to analyze the changes that MSC undergo after administration [8]. Some studies have demonstrated the isolation and reculture of administered MSC from mice [9,10], but none have been able to analyze the function of administered MSC in vivo.

It has become clear that intravenously injected MSC initially accumulate in the lungs due to size restrictions of the lung microvasculature [11]. In the lungs, MSC encounter pulmonary microvascular endothelial cells and resident macrophages and may undergo reciprocal interactions with these cells. It has been demonstrated that MSC affect lung endothelial cells by restoring endothelial permeability by the secretion of hepatocyte growth factor [12].

We have previously demonstrated that the expression levels of multiple cytokines and chemokines in the lungs are modulated after infusion of MSC [13]. The cytokines and chemokines are most likely derived from lung endothelial cells and lung-resident immune cells. Lung-derived factors and intercellular cell surface molecule interactions may have an effect on administered MSC. This would suggest that the functionality of MSC can change already shortly after administration when MSC are present in the lungs and this may alter the therapeutic effect of MSC. A better understanding of the interplay between MSC and the lung tissue-resident cells may lead to optimization of current MSC therapy protocols.

In this study, we examined phenotypical differences between cryopreserved MSC and MSC from continuous culture and analyzed the effect of the lung microvasculature milieu on MSC properties.

Materials and Methods

Isolation and culture of human MSC

MSC were isolated from abdominal subcutaneous adipose tissue of healthy individuals that became available upon kidney donation procedure after written informed consent (protocol no. MEC-2006-190 approved by the Medical Ethics Committee of the Erasmus Medical Center). After collection, the tissue was kept in minimum essential medium-α (MEM-α) (Sigma-Aldrich, St. Louis, MO) supplemented with 1% penicillin/streptomycin solution (P/S; 100 IU/mL penicillin, 100 IU/mL streptomycin; Lonza, Verviers, Belgium) at 4°C and MSC isolated within 24 h. The tissue was minced and enzymatically digested with sterile 0.5 mg/mL collagenase type IV (Life Technologies, Paisley, United Kingdom) at 37°C for 30 min under continuous shaking. The obtained cell suspension was then washed twice, resuspended in culture medium consisting of MEM-α with 1% P/S, 2 mM l-glutamine (Lonza), and 15% fetal bovine serum (FBS; Lonza) and seeded in culture flasks. After 3 days, nonadherent cells were removed. The cultures were kept at 37°C, 5% CO₂, and 95% humidity and the medium refreshed once a week. When the cultures reached 90% confluence, MSC were trypsinized using 0.05% trypsin–ethylenediaminetetraaceticacid (EDTA; Life Technologies, Bleiswijk, Netherlands) and subcultured. MSC were used for experiments at passage 3.

Freeze–thawing procedure

For cryostorage, MSC were removed from their culture flasks by trypsinization and washed in MEM-α with 1% P/S and 15% FBS. They were then resuspended in MEM-α with 1% P/S and 15% FBS at 1 × 10⁶ cells per mL and mixed 1:1 with MEM-α with 20% dimethyl sulfoxide (Merck, Darmstadt, Germany) and 20% FBS, aliquoted in cryovials, and placed in a freezing box (Coolcell; Biocision, San Rafael, CA) at −150°C. To thaw the cells, cryovials were placed in a 37°C water bath until nearly all the ice was melted. The cells were then washed and kept in MEM-α with 1% P/S, 2 mM l-glutamine, and 15% FBS for 1 h at 37°C before they were used in experiments.

Immunophenotyping of MSC

Flow cytometric analysis was conducted on MSC labeled with monoclonal antibodies as outlined in the supporting information (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/scd) and acquired on a FACS Aria (Becton Dickinson, Franklin Lakes, NJ); 2,000–5,000 gated events were quantified and analyzed with Summit v.4.1 software (Dako, Glostrup, Denmark).

In vitro blood clotting assay

The clotting time of human blood was recorded on a semiautomatic 10-channel ball coagulometer (MC10plus; Merlin Medical ABW Medizin und Technik GmbH, Lemgo, Germany), as reported earlier [14]. MSC directly from the culture flask or after thawing were washed twice and resuspended in a buffer containing 5% human serum albumin. Sodium citrate-anticoagulated human blood was obtained from healthy volunteers who had not received any medication for at least 10 days. The cuvette was filled with 100 μL of citrated blood diluted 1:1 in phosphate-buffered saline (PBS). Blood was then supplemented with 50 μL of buffer with or without 3,000 MSC or with 50 μL of positive control reagent. To initiate clotting, 50 μL of 40 mM Ca²⁺ solution was added to a final concentration of 10 mM. The final concentration of MSC was 15,000 cells/mL, corresponding to a dose of 1–2 × 10⁶ cells/kg commonly used in clinical trials.

Administration of human MSC in mice

MSC of four healthy human donors were trypisinized, washed, put through a 40 μm cell strainer, and 1 × 10⁶ MSC in 200 μL PBS were administered in female C57BL/6 mice (Charles River, Wilmington, MA) through tail vein injections. After 2 h, the animals were anesthetized, blood collected in EDTA tubes, and lung tissue removed and snap-frozen for RNA isolation. The animal experiments were approved by the Animal Care and Use Committee of the Erasmus Medical Center (protocol no. EMC-3004).

Human MSC gene expression analysis by RNA sequencing

MSC gene expression was analyzed immediately after removal of MSC from culture flasks, after thawing of cryopreserved MSC, and in MSC that were trapped in the lungs after intravenous administration in mice using mRNA sequencing. For the first group, RNA was isolated from MSC that were snap-frozen immediately after trypsinization (group C). For the second group, RNA was isolated from froze–thawed MSC that were snap-frozen 1 h after recovery at 37°C (group FT). For the third group, RNA was isolated from lung tissue 2 h after infusion of MSC (group I). These samples contained mouse RNA mixed with human RNA from the injected MSC. For each group, MSC of the same three donors of the same passage were used. Lung tissue of a mouse that was injected with PBS was used as a negative control (sample M).

RNA was isolated using Trizol reagent (Life Technologies). Frozen lung tissue was sectioned in 20 μm slices before RNA isolation. Quantity and quality of RNA was assessed using the RNA 6000 Nano kit on a 2100 Bio-analyzer (Agilent, Palo Alto, CA). Samples with an RNA-integrity >8.5 were used. Samples were prepped with TruSeq RNA (v2, Illumina), sequenced SR43 bp on Hiseq2500 in rapid mode, and demultiplexed with CASAVA 1.8.4. Alignment was performed with TopHat 2.0.13 (large index mode, http://tophat.cbcb.umd.edu; with Bowtie 2.2.4.0) against a reference genome containing all human (hg19) and mouse (mm10) chromosomes, with their Refseq gene annotation (Illumina iGenomes, http://support.illumina.com/sequencing/sequencing_software/igenome.html). This custom reference was intended to eliminate cross-alignment from mouse-originating fragments onto their human homologous for the group I samples.

Fragment counts per gene were calculated from the TopHat BAM files, using a custom R (http://r-project.org) script based on the IRanges (Bioconductor; http://bioconductor.org) package and the RefSeq gene annotation. Differential expression analysis was performed with edgeR (Bioconductor), using only the human chromosomes and genes.

Including the mouse chromosomes in the reference genome was not enough to prevent all cross-alignment, as for regions identical between human and mouse, reads get assigned randomly to one of them. This results in human genes having a high read count for sample M, where no alignment against human genes is expected. In samples from group I, this results in highly nonuniform coverage over the transcript for human genes with regions identical between mouse and human. These genes show high peaks in the homologous regions originating from mouse RNA. These genes show up spuriously as significantly differentially expressed. We filtered these genes out of the final list of differentially expressed genes for the comparison of group I versus group C, using a custom-built R script and the dplyr package (CRAN; http://cran.r-project.org). We used two filtering criteria; first, we removed genes that had more than five counts in the negative mouse control sample (M). Second, we removed genes where the counts were localized in less than half of their exons, and the other exons did not show any counts.

Ingenuity software (Qiagen, Venlo, Netherlands) was used for pathway analysis.

Detection of C3 activation fragment a and thrombin–antithrombin complex in murine plasma

Formation of blood activation markers thrombin–antithrombin complex (TAT) and complement component C3 activation fragment a (C3a) in murine plasma was measured with enzyme-linked immunosorbent assay (Cusabio Biotech Ltd., Wuhan, China), at 2, 8, and 24 h post MSC infusion.

Detection of cytokine/chemokine levels in lung tissue

To detect levels of cytokines and chemokines in lung tissue, frozen lung tissue was weighted and sliced in 10-μm-thick slices. The sliced tissue was centrifuged at 15,000 g and tissue fluid collected for cytokine/chemokine measurement by mouse cytokine/chemokine magnetic bead panel multiplex assay (Merck Millipore, Billerica, MA). The panel contained granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony stimulating factor (GM-CSF), interferon (IFN)γ, interleukin (IL)1β, IL6, IL7, IL10, CXCL1, CXCL5, CXCL10, CCL11, monocyte chemoattractant protein 1 (MCP1), MIP1α, and tumor necrosis factor (TNF)α. The samples were measured by Luminex 100/200 cytometer (Luminex, Austin, TX) using Xponent software.

Statistical analysis

Data were tested for significance using analysis of variance and Student's t-test. If the data did not fit a normal distribution, the Mann-Whitney test or the Wilcoxon matched-pairs test was used (two-tailed, 95% confidence intervals). Post hoc analysis was performed using Bonferroni test for multiple comparisons. P-values <0.05 were considered statistically significant.

Results

Gene expression profiling of MSC from continuous culture versus frozen–thawed MSC

To examine the effect of freeze–thawing on MSC, human adipose tissue-derived MSC of three healthy donors from continuous culture (C1, C2, C3) and of frozen-thawed MSC of the same donors (FT1, FT2, FT3) were used for transcriptome profiling by RNA sequencing. Gene expression clustering analysis demonstrated that interdonor gene expression differences were larger than the effects of cryopreservation, as frozen–thawed and cryopreserved MSC clustered per donor (Fig. 1A).

FIG. 1.

mRNA expression analysis of human MSC from continuous culture (C) and frozen–thawed MSC (FT). (A) Clustering of the samples shows that interdonor variation (numbers indicate donors) is larger than the variation between MSC from continuous culture and frozen–thawed MSC (left panel). (B) Heatmap depicting expression patterns of genes significantly different expressed between MSC from continuous culture and frozen-thawed MSC. Data are normalized per row. MSC, mesenchymal stromal cells. Color images available online at www.liebertpub.com/scd

Between the frozen–thawed MSC and continuous culture MSC groups, there were 294 genes that showed a significantly different expression. Of these genes, 167 were upregulated in the frozen–thawed MSC and 127 were downregulated (Fig. 1B). The magnitude of gene expression differences was limited to a maximum increase of 10.9-fold and a maximum decrease of 5.1-fold. A summary of the most significant gene expression changes can be found in Table 1. The full data set can be accessed at GEO accession number GSE76081.

Table 1.

Summary of Gene Expression Differences Between Frozen–Thawed Mesenchymal Stromal Cells Versus Mesenchymal Stromal Cells from Continuous Culture

Gene expression changes of frozen–thawed MSC vs. MSC from continuous culture
Gene name	Fold change up	Gene name	Fold change down
LYPD3	10.9	SMOC2	5.1
Cell–matrix adhesion		Extracellular matrix organization
NRARP	10.9	RGS5	4.7
Negative regulation of Notch signaling		Negative regulation of signal transduction
EGR3	7.0	MIR17HG	3.9
Response to growth factors		Cell survival, proliferation
NPPB	6.2	NEAT1	3.6
Diuretic hormone activity		unknown
RGS2	5.7	CCDC39	3.4
Beta-tubulin binding		Cell motility, movement
RASD1	5.6	C6orf155	3.2
Negative regulation of transcription		Unknown
SNAI1	5.2	MARCH1	3.1
Development, epithelial to mesenchymal transition		Antigen processing, immune response
NR4A3	5.1	CHI3L1	3.0
Apoptosis, proliferation, survival		Inflammatory response
EGR2	5.0	CLDN1	2.8
Development, negative regulation of apoptosis		Cell adhesion, cell–cell junction
NR4A2	4.8	TTC14	2.7
Stress response		Unknown
DUSP5	4.5	SEL1L2	2.7
Activation of MAPK activity		unknown
NR4A1	4.5	MASP2	2.6
Positive regulation of apoptosis and proliferation		Complement activation
DUSP2	4.4	SAA2-SAA4	2.5
Inactivation of MAPK activity		Acute-phase response
ID1	4.0	EVI2B	2.5
Angiogenesis		Unknown
RRAD	3.9	SAA1	2.4
Negative regulation cell growth		Acute-phase response
Pathways differentially expressed and genes involved
MIF regulation of innate immunity↑: z-score 2.2, P < 0.05, 5 of 40 genes different; FOS↑, JUN↑, NFKBIA↑, PLA2G2A↑, PTGS2↑
ILK signaling↑: z-score 3.0, P < 0.05, 10 of 180 genes different; FOS↑, JUN↑, MYC↑, PTGS2↑, RHOB↑, RHOG↑, SNAI1↑, SNAI2↑, TMSB10/TMSB4X↑, VIM↑
Acute phase response signaling↑: z-score 0.82, P < 0.05, 9 of 166 genes different; CARBP2↑, FOS↑, JUN↑, IL6↑, NFKBIA↑, SAA1↓, SAA2↓, SOCS3↑, TNFRSF11B↓
Actin nucleation by ARP-WASP complex↑: z-score 1.34, P < 0.05, 5 of 55 genes different; ARPC1B↑, ITGA2↓, PPP1R12B↓, RHOB↑, RHOG↑
HMGB1 signalling↑: z-score 1.9, P < 0.05, 7 of 117 genes different; FOS↑, JUN↑, IL6↑, IL11↑, RHOB↑, RHOG↑, TNFRSF11B↓

Fifteen genes with the largest increases and 15 genes with the largest decreases are shown. All gene expression changes in the table are significant with P-values <0.05. Brief descriptions of gene functions are indicated underneath the gene names. Five pathways that showed the most significant up or downregulated activity patterns are listed with genes involved.

MSC, mesenchymal stem cells; ↑, upregulation; ↓, downregulation.

Pathway analysis of MSC from continuous culture versus frozen–thawed MSC

A number of individual genes involved in the inhibition of cellular proliferation and induction of growth arrest were upregulated, but pathway analysis did not reveal differential expression of cell survival or cell apoptosis pathways. Freeze–thawing of MSC had an effect on innate immunity pathways (Table 1). These pathways indicated an activation of the acute phase response in frozen-thawed MSC, macrophage inhibitory factor signaling, and activation of Toll-like receptor pathways through high-mobility group protein B1 signalling. A number of genes that are involved in actin rearrangement were upregulated in frozen–thawed cells, such as a number of Rho GTPases and actin-related protein, suggesting active cytoskeletal reorganization processes in MSC after recovering from cryopreservation.

Effects of freeze–thawing on MSC immunophenotype and IBMIR induction

The cell surface expression of a panel of MSC markers was unaffected by freeze–thawing (Fig. 2A). However, a small, but significant increase in the expression of the coagulation factor CD142 (tissue factor, TF) was found on thawed compared to fresh MSC (62% vs. 54% positive). Flow cytometry revealed that higher surface expression of TF on thawed cells went in hand with a small, but significant increase in the number of propidium iodide incorporating cells (84% vs. 91% viable, Fig. 2B). This indicates a small increase in membrane permeability postthawing, which may well explain the increase in TF expression, normally stored in submembrane intracellular granules [2].

FIG. 2.

Effect of freeze–thawing on MSC immunophenotype and triggering of innate immune cascade activation after whole blood exposure in vitro and in vivo. (A) Cell-surface marker expression. N = 7 MSC donors. (B) Cell viability measured by propidium iodide exclusion. (C) MSC from continuous culture and frozen–thawed MSC (15,000 cells per mL) were tested for their triggering of the clotting cascade by exposing them to fresh recalcified human whole blood. Blood clotting time in seconds indicated. Pos, clotting inducing control; 10 mM Ca²⁺ solution. Buffer: negative control; no clotting inducing factors added. (D) Quantification of complement factor C3a (ng/mL) and TAT (ng/mL) in murine plasma after systemic infusion of human MSC (n = 5 MSC donors, two tests each). Bar graphs depict mean ± standard deviation. Dashed lines indicate background levels of C3a and TAT. C3a, C3 activation fragment a; TAT, thrombin–antithrombin complex.

To examine whether freeze–thawing would affect IBMIR triggering, the effect of MSC from continuous culture and frozen–thawed MSC on the clotting time of human blood was examined in vitro. The addition of MSC from continuous culture accelerated blood clotting time (Fig. 2C). Frozen–thawed MSC induced a further, small but significant, acceleration of clotting time compared to continuously cultured MSC.

To corroborate our in vitro findings with the in vivo situation, we analyzed murine EDTA-plasma for formation of complement and coagulation markers at 2, 8, and 24 h post-MSC infusion (Fig. 2D). Infusions of thawed cells lead to significantly higher formation of complement activation marker C3a and coagulation marker TAT at 8 h postinfusion (both P < 0.05).

Effect of MSC from continuous culture and frozen–thawed MSC on lung immune homeostasis

To investigate whether MSC from continuous culture and frozen–thawed MSC had a differential effect on the microenvironment after infusion, 1 × 10⁶ human MSC were injected intravenously in C57BL/6 mice. As MSC accumulate in the lungs after intravenous administration, they may interact with endothelial cells of the lung microvasculature and innate immune cells such as macrophages and granulocytes. Two hours after injection of MSC, lungs were removed and mouse cytokine and chemokine expression analyzed by multiplex assay. Injection of MSC from continuous culture as well as frozen–thawed MSC demonstrated a clear tendency for increased levels of mouse G-CSF, CXCL1, CXCL10, MCP1, and IL6 and a decrease in CXCL5 in lung tissue (Fig. 3). MSC did not induce changes in the expression of IFNγ, CCL11, GM-CSF, IL1β, IL7, IL10, MIP1α, and TNFα (Supplementary Fig. S1). We have previously demonstrated that MSC induce a mild inflammatory response in the lungs, which may be associated with the immunomodulatory effect of MSC [13]. The present data demonstrate that there was no difference between the effects of MSC from continuous culture and frozen–thawed MSC, suggesting that the two cell preparations have a similar immunomodulatory effect in the lungs.

FIG. 3.

Effect of infusion of human MSC from continuous culture (C) and frozen–thawed MSC (FT) on mouse cytokine and chemokine expression in the mouse lung. 1 × 10⁶ MSC were infused through the tail vein and 2 h later, lungs removed for analysis. N = 4 MSC donors. Bar graphs depict mean ± standard deviation. *P < 0.05. G-CSF, granulocyte colony-stimulating factor; MCP1, monocyte chemoattractant protein 1.

Gene expression profiling of MSC after in vivo administration

To examine whether MSC are stabile after in vivo administration or whether they undergo changes under influence of the in vivo environment, we carried out transcriptome analysis on administered MSC. Human MSC from continuous culture (1 × 10⁶) were infused in the tail vein of C57BL/6 mice. The lungs, containing entrapped MSC, were removed 2 h after infusion and total RNA isolated (I1, I2, I3). RNA of all samples was sequenced and mapped against the combined human/mouse genome, thus reflecting human MSC gene expression in the mouse lung. Lung tissue of a mouse that did not receive human MSC was used as a negative control (M). In sample M, positive expression of several human genes was detected, reflecting cross-alignment of mouse RNA on the human genome. These particular genes, which represented 30% of the originally upregulated and 60% of the downregulated genes, were disregarded in samples I1-3. Genes in which <50% of exons showed positive reads in samples I1-3, also reflecting alignment of mouse RNA on the human genome, were also disregarded (20% of the originally differentially expressed genes).

The gene expression profile of the administered MSC differed considerably from both noninjected MSC groups, as can be seen in Figure 4A, where samples C1-3 and FT1-3 now appear to overlap, whereas I1-3 stand out. Negative control sample M stands out from all the other samples (Fig. 4A, B).

FIG. 4.

mRNA expression analysis of human MSC from continuous culture (C), frozen–thawed MSC (FT), and injected MSC (I). (A) Clustering of the samples shows that the effect of the microenvironment on injected MSC is much larger than the effect of freeze–thawing on MSC. Mouse mRNA (M) clearly stands out. (B) Manhattan clustering confirms this. (C) Heatmap depicting expression patterns of genes significantly different expressed between MSC from continuous culture and injected MSC. Data are normalized per row. Color images available online at www.liebertpub.com/scd

Comparison of gene expression profiles between MSC from continuous culture before infusion (C1–3) and after infusion (I1–3) revealed differential expression of 2,060 genes. Of these genes, 720 were upregulated in the infused MSC and 1,340 were downregulated (Fig. 4C). The maximum increase in gene expression in the injected MSC was 1,607-fold. The decreases in gene expression were smaller, with 1,326 of the downregulated genes showing a fold change of <4. The maximum decrease was −10.3-fold. A summary of the most significant gene expression changes can be found in Table 2. The full data set can be accessed at GEO accession number GSE76081.

Table 2.

Summary of Gene Expression Differences Between Injected Mesenchymal Stromal Cells Versus Mesenchymal Stromal Cells from Continuous Culture

Gene expression changes of injected MSC vs. MSC from continuous culture
Gene name	Fold change up	Gene name	Fold change down
NOVA2	1,607	CHAC1	10.3
Regulation of RNA metabolic process		Apoptosis in response to stress
LCE1A	1,585	EID2B Differentiation	7.4
Keratinization		Differentiation
TYRP1	1,252	PPP1R35	6.3
Melanin biosynthetic process		Negative regulation of phosphatase activity
APOC2	1,218	TIGD2	6.1
Cholesterol homeostasis		Unknown
S100B	1,113	CH25H	5.7
Proliferation, differentiation		Lipid metabolic processes
ADRB1	1,075	MMAA	5.1
Adrenergic receptor		Lipid metabolic processes
KRT5	1,075	HDHD3	4.8
Cell junction assembly		Metabolic processes
PRAME	1,046	INE1	4.7
Negative regulation of apoptosis and differentiation		Unknown
SOX17	1,046	KLHDC7B	4.5
Development		Protein ubiquitination
BCL11B	1,003	TTC30B	4.5
Differentiation		Cell projection organization
ZNF366	955	TIGD1	4.2
Response to estrogen stimulus		Unknown
HLA-DRA	815	HILPDA	4.2
Antigen processing and presentation		Positive regulation of cell proliferation, stress response
NRARP	798	GLI4	4.1
Negative regulation of Notch signaling		Regulation of transcription
CD74 (HLA-DR chain)	744	CHAMP1	4.0
Antigen processing and presentation		Protein localization to microtubule
MIR142	657	SERTAD4-AS1 unknown	4.0
Cellular response to estrogen and lipopolysaccharide
Pathways differentially expressed and genes involved
OX40 signaling↓: z-score −0.46, P < 0.05, 12 of 36 genes different; BCL2↑, HLA-DMA↑, HLA-DPA1↑, HLA-DQB1↑, HLA-DRA↑, HLA-DRB1↑, HLA-DRB5↑, MAPK12↓, NFKB1↑, NFKBIB↑, NFKBIE↑
TNF receptor 2 signaling↑: z-score 0.71, P < 0.05, 8 of 26 genes different; BIRC3↑, IKBKG↓, MAP3K14↑, NFKB1↑, NFKBIB↑, NFKBIE↑, TBK1↓, TNFRSF1B↑
Activation of interferon regulatory factor by cytosolic pattern recognition receptors↓: z-score −2.50, P < 0.05, 13 of 59 genes different; IFIH1↑, IFIT2↓, IKBKG↓, IRF3↓, IRF9↓, ISG15↑, MAPK12↓, NFKB1↑, NFKBIB↑, NFKBIE↑, SIKE1↓, STAT2↓, TBK1↓
PKCθ signaling in T lymphocytes↑: z-score 2.45, P < 0.05, 19 of 104 genes different; HLA-DMA↑, HLA-DQA1↑, HLA-DQB1↑, HLA-DRA↑, HLA-DRB1↑, HLA-DRB5↑, IKBKG↓, MAP3K4↑, MAP3K6↓, MAP3K12↓, MAP3K14↑, NFATC1↑, NFATC2↑, NFKB1↑, NFKBIB↑, NFKBIE↑, PIK3C3↓, PIK3CB↑, PIK3R3↑
Induction of apoptosis by HIV1↓: z-score 1.16, P < 0.05, 12 of 56 genes different; BCL2↑, BIRC3↑, DFFA↓, DFFB↓, IKBKG↓, MAP3K14↑, MAPK12↓, NFKB1↑, NFKBIB↑, NFKBIE↑, TNFRSF1B↑, TP53↓

↑, upregulation; ↓, downregulation.

Pathway analysis of MSC after in vivo administration

Pathway analysis 2 h after administration demonstrated that the lung microenvironment affected pathways with an immunological function in MSC. MSC showed strongly increased expression of various human leukocyte antigen class II molecules, which are upregulated in response to inflammatory cytokines. Furthermore, there was a strong upregulation of IFNγ-induced protein 10 (or CXCL10), of the common gamma chain of the receptor for IL2, IL4, IL7, IL9, IL15, and IL21, of nuclear factor of activated T cells, which modulates gene expression during immune activation, and of IFN-regulatory factor 4.

The four pathways that showed the most significantly altered patterns of activation had an immunological function. Immune signaling through the OX40 Tumor Necrosis Factor family pathway and the cytosolic pattern recognition receptor pathway showed reduced activity, whereas the TNF receptor 2 pathway and the phospho-kinase C signaling pathway showed enhanced activity (Table 2), indicating a modulation of the immunomodulatory activity of MSC upon administration.

Even though it is known that MSC have a short survival after intravenous infusion, cell death pathways were not activated. The apoptosis signaling pathway showed no pattern of activation, while the HIV-induced apoptosis pathway showed reduced activity.

Interestingly, multiple genes involved in tyrosine metabolism, such as tyrosinase and tyrosinase-related proteins 1 and 2, were among the most highly upregulated genes.

Production of soluble factors by injected MSC

As the therapeutic effect of MSC is thought to be partly dependent on secreted factors, the expression of genes encoding proteins with an extracellular function was analyzed. Eighteen soluble factors showed upregulated expression of at least 10-fold in injected MSC (Table 3). These factors included factors with an immune function, such as CXCL10, IL11, and IL33, and growth factors and factors that stimulate regeneration by progenitor cells, such as wingless-type MMTV integration site family member (WNTs) and bone morphogenetic protein (BMP)2.

Table 3.

List of Soluble Factors with at Least 10-Fold Upregulated Gene Expression in Injected Mesenchymal Stromal Cells Compared to Noninjected Mesenchymal Stromal Cells

Soluble factors upregulated in injected MSC
Gene name	Fold change up	Function
Apolipoprotein C2	1,261	Increases free fatty acid availability
CXCL10	274	Chemoattraction immune cells
Melanoma inhibitory activity	256	Migration, inhibition of adhesion
WNT7A	239	Development
Neural pentraxin 1	68.6	Innate immunity, acute phase protein
Urocortin 2	64.0	Cardiovascular stimulation
Suppressor of Ty20 homolog like 2	55.7	Nucleic acid-templated transcription
Pleiotrophin	48.5	Growth factor
Apolipoprotein E	42.2	Cholesterol metabolism
Serpin A1	34.3	Protease inhibitor
WAP four-disulfide core domain protein 1	24.3	Protease inhibitor
Parathyroid hormone-like hormone	22.6	Increases Ca²⁺ in the blood
IL11	21.1	Megakaryocyte maturation, osteotrophic
Bone morphogenetic protein 2	18.4	Osteoblast differentiation
IL33	16.0	Drives Th2 immune response
Angiopoietin 4	13.9	Angiogenesis
WNT9A	10.6	Development
Fibroblast growth factor 9	10.6	Growth factor

All gene expression changes listed are highly significant (P < 10⁻⁵, or smaller).

Discussion

The efficiency of MSC therapy will depend for a great part on the phenotype of MSC preparations. In contrast to conventional molecular drugs that, apart from being metabolized by the recipient, do not change, cells can undergo dramatic alterations in response to changes in their microenvironment. Freeze–thawing has been indicated to affect the in vitro immunomodulatory properties of MSC [6,7]. However, MSC functionality is completely recovered after a 24-h culture period. In this study, we found limited gene expression changes in MSC 1 h after thawing. There were changes in genes involved in innate immunity pathways and cytoskeletal rearrangement. The cryopreserved cells were kept in suspension for 1 h after thawing and it is possible that these conditions induced the cells to upregulate cytoskeletal protein expression.

MSC from continuous culture and frozen–thawed MSC induced comparable immunological responses in the mouse lung, the major site of MSC embolization upon intravenous infusion. Indeed, a study by Cruz et al. demonstrated only limited differences between the effects of cryopreserved and continuously cultured MSC in ameliorating airway inflammation, supporting the concept that both MSC preparations are equally effective in vivo [15]. It is likely that the type of model used is determinative for detecting differences between frozen–thawed and continuously cultured MSC or not, as freeze–thawing may affect particular properties of MSC that are employed in particular models, but not in others.

It has previously been demonstrated that islets of Langerhans [16], but also culture expanded MSC, can induce IBMIR [2]. The induction of IBMIR was weakly augmented with frozen-thawed MSC [6,17]. A strong induction of IBMIR would reduce the survival time of MSC after administration and could potentially lead to adverse effects, providing a possible explanation for the limited engraftment of therapeutic MSC [18] and adverse transfusion reactions at higher doses [14]. We found a strong induction of IBMIR for the adipose MSC used in this study, which is in agreement with other reports attributing a strong procoagulant activity to adipose tissue-derived MSC [19 –21]. We measured a weak, but significantly augmented triggering of IBMIR with frozen–thawed cells. The physiological significance of this difference between continuously cultured and frozen–thawed MSC is unclear. However, most importantly, both cell types elicited strong responses. Thus, in vivo persistence will be rather limited with both continuously- and frozen–thawed adipose-derived MSC alike, not necessarily compromising their bioactivity.

In contrast to the minimal effects of freeze–thawing, in vivo administration had a major effect on the transcriptional phenotype of MSC. The observed changes induced in MSC could derive from a number of factors, among them the effects of sheer-stress induced by transport through the bloodstream, factors induced by triggering of IBMIR or the effects of blood cells, and proinflammatory mediators on MSC. MSC accumulate in the lungs upon the first passage [11] and thus within minutes of administration, the majority of MSC are exposed to the microenvironment of the lung microvasculature, where lung endothelial cells and macrophages may undergo interactions with MSC. The gene expression changes in MSC after administration indicate a response to inflammatory signals, eliciting immunological cross talk between entrapped MSC and tissue-resident immune cells. This suggests that cells in the lung microvasculature become activated upon encounter of MSC and respond with inflammatory signals. We have demonstrated previously that lung tissue shows an inflammatory gene expression profile within hours after infusion of MSC [13] and in this study, we demonstrated that levels of the mouse inflammatory factors IL6, CXCL1, CXCL10, and MCP1 were increased in lung tissue upon administration of human MSC.

There is a possibility that the gene expression changes observed in human MSC after infusion in mice are caused by incompatibility of human cells with mouse microenvironment. However, in our earlier study, we found upregulation of immune parameters in mice infused with syngeneic mouse MSC [13]. Interestingly, in this study, we found that human MSC are well capable of responding to mouse cytokines and chemokines. Therefore, we assume that we are not merely looking at a xenoresponse. The nature of the analysis used does, unfortunately, not allow administration of syngeneic cells to test this assumption.

The main question of this study is how the phenotypical changes in MSC affect their function. MSC are known to enhance their immunomodulatory function in response to inflammation [22,23], and we found evidence for increased expression of IL11 and IL33, cytokines of the IL6 and IL1 families, respectively, and for the activation of immune signaling pathways (Fig. 5). There were also nonimmunological pathways that were upregulated in MSC in response to the in vivo environment. There was for instance a dramatic increase in apolipoprotein expression, which is a family of soluble proteins involved in lipid transport. Apolipoproteins may, however, also have anti-inflammatory functions [24]. Furthermore, multiple genes involved in tyrosine metabolism, including tyrosinase, tyrosinase-related protein 1, and dopachrome tautomerase, were strongly upregulated. The significance of these gene expression changes is unknown. MSC showed large increases in WNT7a and WNT9a gene expression. WNT signaling pathways have been identified to play important roles in the control of MSC proliferation and differentiation [25]. Through WNT secretion, MSC may thus target proliferation and differentiation of resident progenitor cells. These gene expression changes indicate that the microenvironment of the lung affects immunological and metabolic activity of MSC. As MSC have a short survival time after infusion, intrinsic changes in, for instance, MSC proliferation rate or differentiation potential are not relevant for the therapeutic effect of MSC. However, changes in cytokine and growth factor secretion by MSC may affect the function of resident cell types, even after MSC have disappeared. In this study, we have not examined the effect of the gene expression changes in MSC on tissues and have not determined whether concentrations of secreted factors reach effective levels. However, we can conclude from the gene expression changes that administered MSC respond rapidly to their new microenvironment and change their immunological and metabolic function.

FIG. 5.

Schematic overview of the changes in MSC 2 h after intravenous administration. Infused MSC end up in the lungs and at 2 h, increase the expression of a range of immunomodulatory and growth factors and change the activity of immune and metabolic pathways. Apo C2, apolipoprotein C2; Apo E, apolipoprotein E; BMP2, bone morphogenic protein 2; FGF9, fibroblast growth factor 9; Wnt7A, wingless-type MMTV integration site family member 7A; Wnt9A, wingless-type MMTV integration site family, member 9A. Color images available online at www.liebertpub.com/scd

Earlier studies have shown that a large majority of intravenously infused MSC disappear within 24 h after administration [9]. We therefore expected to see an upregulation of apoptosis pathways in MSC upon intravenous administration. This was not the case, which would suggest that the disappearance of MSC is independent of the induction of apoptosis. Our data can however not completely prove this. First, apoptosis may be induced independently of gene expression changes. Second, the filter that we used to avoid cross-reactivity of RNA of the mouse lung tissue with the human genome may have filtered out apoptosis genes that are conserved between mouse and human [26]. Therefore, we cannot rule out the possibility that infused MSC undergo apoptosis in the lungs.

Over the last two decades, it has been shown that MSC therapy is safe [27], and the current challenge is to develop efficient therapy. Effort is put in identifying MSC subsets that possess superior immunomodulatory properties and in the development of culture protocols that generate MSC with optimized function. Efficacy testing should allow discrimination between more- and less-potent MSC batches. The findings of this study indicate that MSC show high responsiveness and plasticity upon systemic infusion. Properties that are present in vitro are not necessarily maintained after administration, and vice versa. This is a fact to be taken into account in the development of efficient MSC therapy.

In summary, this study demonstrates that freeze–thawing procedures have little impact on MSC gene expression, but tend to sensitize the cells for stronger recognition by the IBMIR. However, MSC from continuous culture and frozen–thawed MSC had a similar impact on immunological parameters in the lungs. Upon intravenous injection, MSC underwent major gene expression changes, reflecting a response to inflammatory activation. This study describes for the first time that MSC change phenotype and potentially function upon systemic administration, which is important for understanding MSC therapy and improving its efficiency.

Footnotes

Acknowledgment

Guido Moll's contributions were made possible by DFG funding through the Berlin-Brandenburg School for Regenerative Therapies (BSRT, GSC203).

Author Disclosure Statement

No competing financial interests exist.

References

Luk

, de Witte

, Bramer

, Baan

and Hoogduijn

. (2015). Efficacy of immunotherapy with mesenchymal stem cells in man: a systematic review. Expert Rev Clin Immunol, 11:617–636.

Moll

, Rasmusson-Duprez

, von Bahr

, Connolly-Andersen

, Elgue

, Funke

, Hamad

, Lonnies

, Magnusson

, et al. (2012). Are therapeutic human mesenchymal stromal cells compatible with human blood?. Stem Cells, 30:1565–1574.

Galipeau

. (2013). The mesenchymal stromal cells dilemma—does a negative phase III trial of random donor mesenchymal stromal cells in steroid-resistant graft-versus-host disease represent a death knell or a bump in the road?. Cytotherapy, 15:2–8.

Moll

, Jitschin

, von Bahr

, Rasmusson-Duprez

, Sundberg

, Lonnies

, Elgue

, Nilsson-Ekdahl

, Mougiakakos

, et al. (2011). Mesenchymal stromal cells engage complement and complement receptor bearing innate effector cells to modulate immune responses. PLoS ONE, 6:e21703.

Doorn

, Moll

, Le Blanc

, van Blitterswijk

and de Boer

. (2012). Therapeutic applications of mesenchymal stromal cells: paracrine effects and potential improvements. Tissue Eng Part B Rev, 18:101–115.

Moll

, Alm

, Davies

, von Bahr

, Heldring

, Stenbeck-Funke

, Hamad

, Hinsch

, Ignatowicz

, et al. (2014). Do cryopreserved mesenchymal stromal cells display impaired immunomodulatory and therapeutic properties?. Stem Cells, 32:2430–2442.

Francois

, Copland

, Yuan

, Romieu-Mourez

, Waller

and Galipeau

. (2012). Cryopreserved mesenchymal stromal cells display impaired immunosuppressive properties as a result of heat-shock response and impaired interferon-(gamma) licensing. Cytotherapy, 14:147–152.

Karp

and Leng Teo

. (2009). Mesenchymal stem cell homing: the devil is in the details. Cell Stem Cell, 4:206–216.

Eggenhofer

, Benseler

, Kroemer

, Popp

, Geissler

, Schlitt

, Baan

, Dahlke

and Hoogduijn

. (2012). Mesenchymal stem cells are short-lived and do not migrate beyond the lungs after intravenous infusion. Front Immunol, 3:297.

10.

Belema-Bedada

, Uchida

, Martire

, Kostin

and Braun

. (2008). Efficient homing of multipotent adult mesenchymal stem cells depends on FROUNT-mediated clustering of CCR2. Cell Stem Cell, 2:566–575.

11.

Fischer

, Harting

, Jimenez

, Monzon-Posadas

, Xue

, Savitz

, Laine

and Cox

. (2009). Pulmonary passage is a major obstacle for intravenous stem cell delivery: the pulmonary first-pass effect. Stem Cells Dev, 18:683–691.

12.

Chen

, Liu

, Qiu

and Yang

. (2015). Interaction between mesenchymal stem cells and endothelial cells restores endothelial permeability via paracrine hepatocyte growth factor in vitro. Stem Cell Res Ther, 6:44.

13.

Hoogduijn

, Roemeling-van Rhijn

, Engela

, Korevaar

, Mensah

, Franquesa

, de Bruin

, Betjes

, Weimar

and Baan

. (2013). Mesenchymal stem cells induce an inflammatory response after intravenous infusion. Stem Cells Dev, 22:2825–2835.

14.

Moll

, Ignatowicz

, Catar

, Luecht

, Sadeghi

, Hamad

, Jungebluth

, Dragun

, Schmidtchen

and Ringden

. (2015). Different procoagulant activity of therapeutic mesenchymal stromal cells derived from bone marrow and placental decidua. Stem Cells Dev, 24:2269–2279.

15.

Cruz

, Borg

, Goodwin

, Sokocevic

, Wagner

, McKenna

, Rocco

and Weiss

. (2015). Freshly thawed and continuously cultured human bone marrow-derived mesenchymal stromal cells comparably ameliorate allergic airways inflammation in immunocompetent mice. Stem Cells Transl Med, 4:615–624.

16.

Moberg

, Johansson

, Lukinius

, Berne

, Foss

, Kallen

, Ostraat

, Salmela

, Tibell

, et al. (2002). Production of tissue factor by pancreatic islet cells as a trigger of detrimental thrombotic reactions in clinical islet transplantation. Lancet, 360:2039–2045.

17.

Moll

, Hult

, von Bahr

, Alm

, Heldring

, Hamad

, Stenbeck-Funke

, Larsson

, Teramura

, et al. (2014). Do ABO blood group antigens hamper the therapeutic efficacy of mesenchymal stromal cells?. PLoS One, 9:e85040.

18.

von Bahr

, Batsis

, Moll

, Hagg

, Szakos

, Sundberg

, Uzunel

, Ringden

and Le Blanc

. (2012). Analysis of tissues following mesenchymal stromal cell therapy in humans indicates limited long-term engraftment and no ectopic tissue formation. Stem Cells, 30:1575–1578.

19.

Tatsumi

, Ohashi

, Matsubara

, Kohori

, Ohno

, Kakidachi

, Horii

, Kanegae

, Utoh

, Iwata

and Okano

. (2013). Tissue factor triggers procoagulation in transplanted mesenchymal stem cells leading to thromboembolism. Biochem Biophys Res Commun, 431:203–209.

20.

Shiratsuki

, Terai

, Murata

, Takami

, Yamamoto

, Fujisawa

, Burganova

, Quintanilha

and Sakaida

. (2015). Enhanced survival of mice infused with bone marrow-derived as compared with adipose-derived mesenchymal stem cells. Hepatol Res, 45:1353–1359.

21.

Jung

, Kwon

, Choi

, Shin

, Park

, Choi

and Kim

. (2013). Familial occurrence of pulmonary embolism after intravenous, adipose tissue-derived stem cell therapy. Yonsei Med J, 54:1293–1296.

22.

Crop

, Baan

, Korevaar

, Ijzermans

, Pescatori

, Stubbs

, van Ijcken

, Dahlke

, Eggenhofer

, Weimar

and Hoogduijn

. (2010). Inflammatory conditions affect gene expression and function of human adipose tissue-derived mesenchymal stem cells. Clin Exp Immunol, 162:474–486.

23.

Krampera

, Cosmi

, Angeli

, Pasini

, Liotta

, Andreini

, Santarlasci

, Mazzinghi

, Pizzolo

, et al. (2006). Role for interferon-gamma in the immunomodulatory activity of human bone marrow mesenchymal stem cells. Stem Cells, 24:386–398.

24.

White

, Garber

and Anantharamaiah

. (2014). Anti-inflammatory and cholesterol-reducing properties of apolipoprotein mimetics: a review. J Lipid Res, 55:2007–2021.

25.

Ling

, Nurcombe

and Cool

. (2009). Wnt signaling controls the fate of mesenchymal stem cells. Gene, 433:1–7.

26.

Huang

, Deveraux

, Maeda

, Salvesen

, Stennicke

, Hammock

and Reed

. (2000). Evolutionary conservation of apoptosis mechanisms: lepidopteran and baculoviral inhibitor of apoptosis proteins are inhibitors of mammalian caspase-9. Proc Natl Acad Sci U S A, 97:1427–1432.

27.

Lalu

, McIntyre

, Pugliese

, Fergusson

, Winston

, Marshall

, Granton

, Stewart

and Canadian Critical Care Trials Group. (2012). Safety of cell therapy with mesenchymal stromal cells (SafeCell): a systematic review and meta-analysis of clinical trials. PLoS One, 7:e47559.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.05 MB

0.02 MB