Sorting living mesenchymal stem cells using a TWIST1 RNA-based probe depends on incubation time and uptake capacity
- 222 Downloads
Bone marrow derived mesenchymal stromal cells (BMSCs) are multipotent progenitors of particular interest for cell-based tissue engineering therapies. However, one disadvantage that limit their clinical use is their heterogeneity. In the last decades a great effort was made to select BMSC subpopulations based on cell surface markers, however there is still no general consensus on which markers to use to obtain the best BMSCs for tissue regeneration. Looking for alternatives we decided to focus on a probe-based method to detect intracellular mRNA in living cells, the SmartFlare technology. This technology does not require fixation of the cells and allows us to sort living cells based on gene expression into functionally different populations. However, since the technology is available it is debated whether the probes specifically recognize their target mRNAs. We validated the TWIST1 probe and demonstrated that it specifically recognizes TWIST1 in BMSCs. However, differences in probe concentration, incubation time and cellular uptake can strongly influence signal specificity. In addition we found that TWIST1high expressing cells have an increased expansion rate compared to TWIST1low expressing cells derived from the same initial population of BMSCs. The SmartFlare probes recognize their target gene, however for each probe and cell type validation of the protocol is necessary.
KeywordsMesenchymal stem cells (MSCs) TWIST1 RNA probes Tissue engineering Cell sorting Expansion
Multipotent progenitor cells from bone marrow aspirates can differentiate into chondrocytes, osteoblasts and adipocytes (Pittenger et al. 1999). These progenitor cells, often referred to as bone marrow-derived mesenchymal stem or stromal cells (BMSCs), are appealing for cell-based tissue engineering purposes. Unfortunately, their limited expansion capacity and their heterogeneity, hinder their clinical use (Banfi et al. 2000; Bonab et al. 2006; Chen et al. 2005; Li et al. 2011). Several studies investigated cell surface molecules to identify specific subpopulations of BMSCs (Alvarez-Viejo et al. 2015; Buhring et al. 2007; Cleary et al. 2016; Delorme et al. 2008; Sacchetti et al. 2007; Sivasubramaniyan et al. 2013). However, despite the great effort, there is still no general consensus on the surface markers that need to be used to define or select the best BMSC subset for tissue engineering. One drawback of surface markers is that their function is often unknown, so alternative markers are necessary to select cells according to their function (Clevers and Watt 2018).
Recently, a probe-based method to detect intracellular mRNA in living single cells has been developed, the SmartFlare technology (Seferos et al. 2007; Prigodich et al. 2009). The SmartFlare technique is a promising tool to sort BMSCs into functionally different populations. The SmartFlare probes are taken up by the cells via endocytosis and if the target mRNA is present, the probes bind to the target mRNA and fluorescent reporters are released and detectable (Figure S1A). Since the SmartFlare technology is available, this technique already successfully identified cancer cells (McClellan et al. 2015; Kronig et al. 2015) and pluripotent stem cells (Lahm et al. 2015). Additionally it was applied to investigate a Nodal expressing subpopulation of melanoma cells (Seftor et al. 2014), and to study a subpopulation of human BMSCs with an enhanced osteogenic potential (Li et al. 2016). However, other studies did not find a correlation between the SmartFlare fluorescence intensity and mRNA expression measured by RT-PCR (Czarnek and Bereta 2017; Yang et al. 2018). In addition, Czarnek et al. (2017) found that the SmartFlare signal intensity correlates with the probe uptake ability of the cells.
To assess if the SmartFlare technique can be used to sort different populations of BMSCs based on gene expression, we focused on the validation of a probe for TWIST1. TWIST1 is a transcription factor that is involved in the regulation of BMSC proliferation (Goodnough et al. 2012; Isenmann et al. 2009; Tian et al. 2015) and differentiation (Isenmann et al. 2009; Boregowda et al. 2016; Cleary et al. 2017; Narcisi et al. 2015, 2016). In the present study, we evaluated the SmartFlare protocol in order to detect a specific probe signal in our culture conditions and illustrated that the SmartFlare fluorescence intensity is associated with probe concentration, incubation time and cellular uptake capacity.
Materials and methods
Isolation and culture of human adult bone marrow mesenchymal stem cells
Human adult bone marrow aspirates were obtained from femoral biopsies of 8 patients (22–79 years) undergoing total hip replacement (MEC 2015-644, MEC 2004-142: Erasmus Medical Center, Rotterdam; MEC 2011.07 Albert Schweitzer Hospital, Dordrecht), after obtaining informed consent and full ethical approval by the Erasmus MC and Albert Schweitzer ethics committee.
Human BMSCs were isolated, seeded at the density of 2300 cells/cm2 and cultured as previously described in standard expansion media, containing 10% FCS (Lonza, Verviers, Belgium; selected batch:1S016) and 1 ng/ml FGF2 (AbD Serotech, Kidlington, United Kingdom) (Narcisi et al. 2016). The medium was refreshed twice a week. BMSCs expanded for 3 to 6 passages were used for experiments.
Cells were treated with the SmartFlare probe when they were sub-confluent. SmartFlare probes TWIST1-Cy3 (the only label available for TWIST1), Uptake-Cy5, and GAPDH-Cy5 were purchased from Merck. The probes were resuspended in 50 µl sterile nuclease free water, 1:20 prediluted in PBS (Lonza) and added to the cells with a final concentration of 50 pM or 100 pM. The cells were incubated for 6 or 16 h at 37 °C and 5% CO2 and analyzed using flow cytometry. To assure a broad range of TWIST1 gene expression during the validation of the TWIST1-Cy3 probe, BMSCs from two different donors were mixed and treated with the TWIST1-Cy3 probe.
Flow cytometry and FACS
Flow cytometry analysis was performed using a BD Fortessa and the data was analyzed using FlowJo V10 software. The cells were sorted using a BD Biosciences FACS Aria and the data was analyzed using BD FACS Diva 8.0.1 software. Cell debris were excluded from the population through forward scatter (FSC)/side scatter (SSC) gate and doublets were excluded using FSC-A/FSC-H gate (Figure S2A). To confirm effective sorting, the sorted populations were reanalyzed (Figure S2B). Mean fluorescent intensity (MFI) was measured using FlowJo V10 software. The two different gates TWIST1high and TWISTlow were established based on the TWIST1-Cy3 fluorescence intensity, 15–25% of the extremes or two different gates TWIST1/Uptakehigh and TWIST1/Uptakelow were established based on the TWIST1-Cy3 fluorescence intensity, 15% of the extremes with a comparable Uptake-Cy5 fluorescence intensity. The sorted cells were collected in PBS with 1% FCS and reseeded with a density of 2300 cells per cm2 or used for RNA isolation.
Real time PCR analysis
Post-sorting, 200,000 BMSCs per sample were spun down and treated on ice with RLT lysis buffer (Qiagen, Hilden, Germany) with 1% β-mercaptoethanol. BMSCs in monolayer were washed with PBS and treated on ice with RLT lysis buffer (Qiagen) with 1% β-mercaptoethanol. A range of 0.25–1.00 µg of purified RNA (RNeasy Micro Kit; Qiagen) was reverse transcribed into cDNA (RevertAid First Strand cDNA Synthesis Kit; MBI Fermentas, St. Leon-Rot, Germany). RT-PCR was performed using an annealing temperature of 60 °C on a C1000 Touch™ Thermal Cycler using SybrGreen (Eurogentec, Seraing, Belgium). The data were normalized to the housekeeper gene RSP27a. The relative expression was calculated according to the 2−ΔΔCt formula. The primers used for RT-PCR are listed in (Table S1).
Linear correlation (Fig. 2c) was analyzed with GraphPad Prism Software 5.00 assuming normal distribution of the data.
TWIST1 SmartFlare detects TWIST1 mRNA after 6 h using a concentration of 50 pM in human BMSCs
To study TWIST1-Cy3 signal specificity, BMSCs were treated with TWIST1-Cy3 probe for 16 h or 6 h, sorted based on the TWIST1-Cy3 signal by FACS and subsequently tested by RT-PCR. Our FACS gating strategy consisted of sorting 15% of the BMSCs with the lowest TWIST1-Cy3 signal and 15% of the BMSCs with the highest TWIST1-Cy3 signal (TWIST1low vs TWIST1high; Fig. 1b). To our surprise no difference in relative TWIST1 gene expression was detected between TWIST1low and TWIST1high cells after 16 h of probe incubation (Fig. 1c). This indicates that although we observe a TWIST1 SmartFlare signal after 16 h, this signal is probably not specific for TWIST1 gene expression. However after 6 h incubation we confirmed that TWIST1high BMSCs have a higher TWIST1 gene expression than the TWIST1low population (6.25-fold difference; Fig. 1c). These data shows that the TWIST1 probe specifically detects TWIST1 gene expression in this population of BMSCs already after 6 h incubation with a concentration of 50 pM probe. In addition we observed that more than 97.3% of cells were positive for the Uptake control probe, a probe which is always fluorescent without binding to a target (Supplementary Figure 1B), with 50 pM after 6 h of incubation.
Correction for cellular probe uptake improves TWIST1 gene detection
When we repeated the sorting experiment with other donors not always differences in TWIST1 expression by RT-PCR were observed between TWIST1low and TWIST1high sorted cells (Figure S4). Given that, and considering that Czarnek et al. recently showed that uptake capacity can influence the SmartFlare signal specificity (Czarnek and Bereta 2017), we decided to carefully monitor uptake in our BMSC populations.
TWIST1high BMSCs have a high expansion capacity
In this study, we evaluated the use of the SmartFlare technique to detect TWIST1 expression at a single cell level in living BMSCs. Multiple studies successfully detected mRNA expression with the SmartFlare technique (McClellan et al. 2015; Kronig et al. 2015; Lahm et al. 2015; Seftor et al. 2014; Li et al. 2016). However, two recent studies showed that different SmartFlare probes were not able to specifically detect their target mRNAs in cell lines and monocytes (Czarnek and Bereta 2017; Yang et al. 2018). Here we showed that SmartFlare is an effective tool to detect TWIST1 gene expression in living BMSCs, but differences in probe concentration, incubation time and cellular uptake can influence the SmartFlare sensitivity and possibly lead to misinterpretation of the results.
We observed that specific detection of TWIST1 mRNA expression in BMSCs is possible already after 6 h of incubation with a concentration of 50 pM, TWIST1-Cy3 probe. While most of the studies used 16 h (McClellan et al. 2015; Seftor et al. 2014; Li et al. 2016; Czarnek and Bereta 2017) or even a longer incubation time (Kronig et al. 2015; Lahm et al. 2015; Czarnek and Bereta 2017), we were not able to specifically detect TWIST1 after 16 h incubation (Figure S4). The SmartFlare technology was recently applied in BMSCs (Li et al. 2016), but never for the detection of TWIST1 expression. In our study a different protocol was needed compared to the RUNX2 and the SOX9 probes used by Li et al. (2016). Possible explanations could be ascribed to differences in culture conditions, origin of BMSCs or binding efficiency of the probe to the target.
In addition, our data indicate that BMSCs can have a high difference in probe uptake. We observed that these differences strongly influence the TWIST1 SmartFlare specificity. This confirms confirms the data previously reported where was shown that SmartFlare intensity was affected by cellular uptake in 293T cells (Czarnek and Bereta 2017). The differences in uptake capacity can be explained by differences in cell cycle stage between the BMSCs, since endocytosis is reduced during mitosis (Fielding et al. 2012). Here, we were able to overcome this problem by correcting TWIST1 detection for the cellular uptake based on Uptake probe intensity during sorting. Next, we demonstrate that TWIST1high expressing BMSCs have a higher expansion capacity than TWIST1low expressing BMSCs derived from the same donor. A population of BMSCs with a high TWIST1 expression and a high proliferation rate have already been reported by us and others (Isenmann et al. 2009; Cleary et al. 2017; Narcisi et al. 2015). Here, we showed for the first time that within the same population of BMSCs, the subpopulation of TWIST1high expressing cells have a higher expansion capacity than the TWIST1low expressing cells. Alternatively to the use of the uptake control, in a previous report the ratio between two functional markers, RUNX2 and SOX9, was applied (Li et al. 2016). This indirect method could also be used, since it would automatically take into account differences in uptake, as these would not change the ratio, but only the intensity of the individual signals.
In summary, our data indicate that for each probe and cell type, a validation of the SmartFlare protocol is necessary. Giving that, we were able to successfully use the TWIST1 probe to detect TWIST1 mRNA in living BMSCs and to sort TWIST1high BMSCs from a heterogeneous population of cells. Overall, we showed that SmartFlare is a promising tool to divide a heterogeneous population of cells based on gene expression in functionally different populations.
J.L.M. Koevoet, A.L.M. Vriends, T.W. Kan, P. van Geel, G. Kremers are acknowledged for technical assistance. We are thankful to the Erasmus MC FACS sorting facility for support with FACS machines This work was financially supported by the Dutch Arthritis Foundation (ReumaNederland; 16-1-201), and R.N. was further supported by a Veni grant from NWO (13659).
Compliance with ethical standards
Conflict of interest
The authors declare no competing interests.
Ethics approval and consent to participate
Human adult bone marrow aspirates were obtaine after obtaining informed consent and full ethical approval by the Erasmus MC and Albert Schweitzer ethics committee.
- Lahm H, Doppler S, Dressen M, Werner A, Adamczyk K, Schrambke D, Brade T, Laugwitz KL, Deutsch MA, Schiemann M, Lange R, Moretti A, Krane M (2015) Live fluorescent RNA-based detection of pluripotency gene expression in embryonic and induced pluripotent stem cells of different species. Stem Cells 33:392CrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.