Gene amplification during differentiation of mesenchymal stem cells towards chondrocytes

Introduction

Gene amplification is a genetic process where the copy number of certain genes increases, without a proportional increase of the genome. Gene amplification is not only a hallmark of genetic alterations in tumor cells but is also well described and studied since decades as physiological mechanism in amphibians and flies. For example, during the development of Drosophila melanogaster and Xenopus laevis, specific DNA regions are amplified within defined developmental time windows (1).

There is growing evidence that gene amplification plays a physiological role during differentiation of stem and progenitor cells. Previously we described gene amplification during differentiation of mesenchymal stem cells towards osteoblasts and adipocytes (2). Interestingly, gene amplifications that arise during stem cell differentiation often overlap with those found in the corresponding tumor types. For example, the genes CDK4 and MDM2 were shown to undergo co-amplification during the differentiation of mesenchymal stem cells (MSCs) into adipocytes, mirroring the high-level amplifications characteristic of liposarcoma (3). Similarly, several studies on chondrosarcoma have reported MYC gene amplification on chromosome 8, with array-CGH analyses demonstrating that MYC amplification occurs more frequently in higher-grade and dedifferentiated chondrosarcomas (4,5). In addition, genomic profiling of chondrosarcoma revealed several amplified chromosome regions including regions on chromosomes 5, 12, 19 and 22(6). Notably, amplification of CDK4 without accompanying MDM2 co-amplification has also been documented in at least one chondrosarcoma case, further underscoring the heterogeneity and potential clinical relevance of these genetic events (7).

Mesenchymal stem cells (MSCs) possess significant potential for therapeutic applications in tissue regeneration. One of the most ambitious investigations is probably regeneration of articular cartilage since many health restrictions are based on cartilage destruction (8-10). Understanding both the regenerative capacity and the potential risks caused by genetic changes that are associated with directing hMSCs toward a chondrogenic lineage is therefore essential. Here we set out to investigate whether gene amplifications are detectable during chondrogenic differentiation of human mesenchymal stem cells (hMSCs).

Materials and methods

Cell culture and differentiation

Primary hMSCs from bone marrow were purchased form PromoCell and were cultured until they formed spheres in U-bottom 96-well plates or were near confluent on Zeiss FrameSlides PET (polyethylene terephthalate) membrane for laser microdissection. Chondrocyte differentiation was induced in conditions using chondrogenic differentiation medium (PromoCell) (11) for 5-7 days.

A detailed characterization of hMSCs is provided in the supplier's certificates of analysis for Lot numbers 488Z010 (female donor, 79 years) and 479Z025 (male donor, 80 years), both of which were used in this study. As indicated in these certificates, surface marker analysis at passage 3 showed that 100% of cells were positive for CD73, CD90 and CD105, while 99-100% were negative for CD34 CD45. The certificates also document differentiation potential toward adipogenic, osteogenic, and chondrogenic lineages. All experiments were performed using cells at passage 3.

RNA-seq

RNA was isolated using miRNAeasy Kit (Qiagen) and used for next generation sequencing. Sequencing was performed on an DNBSEQ-G400RS instrument by the NGS Sequencing Facility of Saarland University using the 100bp paired end sequencing strategy. Paired-end fastq files were processed using the mRNA module of SnakePipes 3.0.0(12). The reads were aligned using STAR using default settings provided by snakepipes (13) against GRCh38(14) in order to generate bam and bai files. Quality control checks were carried out using multiQC (15).

Toluidine blue staining of proteoglycan during differentiation

HMSCs were cultivated and differentiated as described above. For toluidine blue staining cells were washed once with PBSx1 for 10 min and either fixed with 4% PFA (paraformaldehyde) for 5 min or with ice cold methanol for 10 min. Toluidine staining solution (10% Toluidine blue, 1% Borax) was applied for 15 min, and excess staining was removed by three PBSx1 washing steps (16). Differentiation was checked with a bright-field microscope at x20 magnification.

Laser microdissection of differentially toluidine-blue stained areas

For laser microdissection hMSCs were differentiated for 5 days towards chondrocytes and stained with toluidine blue for proteoglycan detection of differentiating cells. In total 20 areas (for each staining subpopulation: grey, light blue, blue and dark blue) of 0.76 mm2 were laser microdissected, using the PALM® MicroBeam Microdissection System (Carl Zeiss AG). These experiments were done in three biological replicates. DNA extraction was performed using the QIAamp DNA Micro Kit (Qiagen) according to the manufacturer's instructions. The final eluate was concentrated to a total volume of approximately 12 µl and DNA concentration was determined using Qubit™ dsDNA HS Assay Kit (Thermo Fisher Scientific).

qPCR analysis

TaqMan Copy Number Assays for genes CDK4 (Hs00957586_cn), MDM2 (Hs00181272_cn), AGAP2 (Hs01425078_cn), CPT1B (Hs01741692cn), MYC (Hs02602824_cn), SHANK3 (Hs04081743_cn), and TRIB1 (Hs00836770_cn) were performed following manufacturer's instructions (Applied Biosystems®). RNaseP TaqMan Copy number reference assay was used for relative quantitation of copy number of target genes. DNA from not differentiated controls of hMSCs cultivated in DMEM was used as control standard for normal diploid copy number. TaqMan assays were run in four technical replicates and results were analyzed using StepOne™ Software v2.0 and CopyCaller™ software (Applied Biosystems® by Life Technologies Corporation) for differentiation in U-bottom 96-well plates. TaqMan Copy Number Assays for genes CDK4 (Hs00957586_cn), AGAP2 (Hs01425078_cn), CPT1B (Hs01741692cn), SHANK3 (Hs04081743_cn), and TRIB1 (Hs00836770_cn) were performed with 250 pg DNA per reaction. TaqMan assays were run in two technical replicates and results were analyzed using StepOne™ Software v2.0 and CopyCaller™ software for laser microdissected areas. For analysis of this low amount of DNA input the Cq-threshold was set to exclude values above 35.

Statistical analysis

Copy number values extracted from CopyCaller™ were analyzed for each gene separately. One-way ANOVA (alpha level 0.05, n=6) was used to test the null hypothesis that staining color (CTRL, grey, light blue, blue, dark blue) and copy number are independent for the genes AGAP2, CDK4, CPT1B, SHANK3 and TRIB1. We performed Tukey's HSD post hoc test (alpha=0.05) to find pairs of staining area that are significantly different in terms of copy number for each of the genes mentioned. Calculations were performed using R version 4.3.3, specifically the aov function (https://search.r-project.org/R/refmans/stats/html/aov.html) with the formula Copy_number ~ Staining followed by the Tukey HSD function for post hoc comparisons. P<0.05 was considered to indicate a statistically significant difference.

Results

HMSCs were differentiated in U-bottom 96-wells and differentiation was confirmed by toluidine blue staining and RNA-seq analysis of chondrogenic differentiation specific genes SOX9, COL2A1 and ACAN (17,18). Already at 5 days post differentiation induction cells revealed three-dimensional cell growth as spheres and high intensity of proteoglycan specific toluidine blue staining (Fig. 1). Expression analysis of two independent biological replicates revealed an increase of SOX9 expression after 2 days and a further increase after 7 days of differentiation and an increase of COL2A1 and ACAN expression after 5 and 7 days of differentiation (Fig. 2). To detect gene amplifications, we analyzed bulk DNA from hMSCs differentiated toward chondrocytes for 1, 2, 5, and 7 days, as well as from undifferentiated hMSCs serving as controls. Amplification was assessed by quantitative PCR (TaqMan) in four replicates, and data were evaluated using the CopyCaller software (Applied Biosystems).

Figure 1 Proteoglycan and spheroid formation during chondrocyte differentiation. Human mesenchymal stem cells were differentiated in U-bottom wells, and spheroid formation together with proteoglycan deposition using toluidine blue staining was monitored over 14 days. (A) Proteoglycan staining by toluidine blue was detectable as early as day 2 in the absence of spheroid formation. (B) Spheroid formation became evident by day 5. Strong toluidine blue staining was observed in all spheroids at days (C) 7 and (D) 14. (E) An unstained spheroid at day 14 is shown as a reference. Images were taken at 20x magnification.

Figure 2 Expression of SOX9, COL2A, and ACAN during chondrocyte differentiation. RNA-seq data from two biological replicates (shown as 1. and 2. on the left) are visualized using Integrative Genomics Viewer. (A) Undifferentiated hMSCs exhibited minimal SOX9 expression; levels increased at days 2 and 5 and reached a maximum at day 7, consistent with chondrocyte differentiation. (B) COL2A1 and (C) ACAN expression were undetectable at day 2 but became apparent at day 5 and peaked at day 7, further confirming successful chondrogenic differentiation. X-axis indicates chromosome localization in Mb and y-axis RNA-seq results represented as read-height.

For our investigation we selected genes previously found to be amplified during osteoblast and adipocyte differentiation and genes that were known to be amplified in tumors derived from chondrocytes including genes on chromosome 12, chromosome 22 and chromosome 8 (2,6). Copy number was determined for genes AGAP2, CDK4 and MDM2 localized on chromosome 12, CPT1B and SHANK3 localized on chromosome 22 and TRIB1 and MYC localized on chromosome 8. As shown in Fig. 3 after 1 day of differentiation none of the investigated genes revealed a copy number increase. After 2 days of differentiation CDK4, CPT1B and SHANK 3 revealed a 60% copy number increase, AGAP2 and TRIB1 exhibited a 20% copy number increase; MDM2 and MYC remained unchanged (Fig. 3). After 5 days of differentiation AGAP2 and MDM2 displayed increases, whereas CDK4, CPT1B and SHANK3, which were highest at 2 days, decreased and MYC still remained unchanged. After 7 days of differentiation CDK4, CPT1B, SHANK3, and TRIB1 decreased further, AGAP2 increased, and no further copy number change was detected for MDM2. MYC showed no change throughout the 7-day differentiation period.

Figure 3 Descriptive amplification analysis of AGAP2, CDK4, MDM2, CPT1B, SHANK3, TRIB1 and MYC using qPCR during chondrogenic differentiation. Amplification of AGAP2, CDK4, MDM2, CPT1B, SHANK3, TRIB1 and MYC was analyzed by qPCR using TaqMan copy number assays and RNAseP was used as reference TaqMan assay. DNA from undifferentiated hMSCs (CTRL) served as standard for normal diploid copy number with copy number values indicated on the y-axis and normal diploid copy number is referred as 2. Copy numbers are shown as mean from four technical replicates from one experiment (3 pooled spheroid samples) with vertical lines indicating the range. Chondrogenic-induced hMSC cells were analyzed at four time points after differentiation induction (1, 2, 5 and 7 days). CPT1B and SHANK3 revealed the highest copy numbers of all investigations after 2 days of differentiation and a decrease until 7 days. CDK4 revealed also a high copy number after 2 days of differentiation induction (green) and a decrease until 7 days. TRIB1 revealed a copy number increase only 2 days after differentiation induction. AGAP2 and MDM2 revealed a copy number increase only after 5 and 7 days. MYC revealed no copy number increase at any time point after differentiation induction. qPCR, quantitative PCR.

Previous published investigations on gene amplification during differentiation revealed a high heterogeneity of gene amplification between single cells. During osteoblast differentiation for example, cells with high level amplification were detected next to cells with normal copy number and during neural differentiation composition and copy number of co-amplified genes varied between neighboring cells (2,19). One possible reason for the differences in gene copy numbers is differences in the degree of differentiation upon differentiation induction. After five days upon differentiation induction, we performed laser microdissection on hMSCs after toluidine blue staining for detection of proteoglycan production that is indicative for chondrocyte differentiation (Fig. 4A-D). Toluidine blue stain revealed heterogenous blue intensities with areas of grey (Fig. 4A), light blue (Fig. 4B), blue (Fig. 4C) and dark blue (Fig. 4D) staining as result of different amounts of produced proteoglycans indicating different stages of differentiation.

Figure 4 Amplification analysis of AGAP2, CDK4, CPT1B, SHANK3 and TRIB1 after laser microdissection. hMSCs were differentiated towards chondrocytes for 5 days on PET frame slides for laser microdissection. Cells were seeded at high density, and slide was stained with toluidine blue for proteoglycan detection during differentiation. Different areas of staining intensity could be detected and were defined as areas with (A) grey cells, (B) light blue cells, (C) blue cells and (D) dark blue cells, all indicating stages during the differentiation towards chondrocytes. In total for each area type twice 20 areas were laser microdissected and used for DNA isolation. Experiments were done in three biological replicates. Examples of laser microdissected areas are circled in white and scale bars represent 150 µm. (E) Results of amplification analysis of our quantitative PCR using TaqMan copy number assays and RNAseP as reference are shown as diagram with AGAP2, CDK4 and CPT1B SHANK3 and TRIB1. DNA from undifferentiated hMSCs (CTRL) served as standard for normal diploid copy number with copy number values indicated on the y-axis and normal diploid copy number is referred as 2. Copy numbers are presented as means of two technical replicates across three biological replicates (N=6), with error bars representing standard deviation. *P<0.05. CDK4 revealed the highest copy number in areas with grey and light blue cells indicating gene amplification at the beginning of differentiation. CPT1B revealed high copy numbers in grey, light blue and blue cells indicating gene amplification during differentiation and SHANK3 revealed the highest copy numbers in areas with blues cells indicating gene amplifications during later differentiation progression. AGAP2 and TRIB1 revealed no copy number increase. In areas with dark blues cells no significant amplification was detectable indicating a reduction of gene amplification towards final differentiation. hMSC, human mesenchymal stem cell.

In total we laser microdissected 20 areas for each blue intensity as shown in Fig. 4A-D. Although the laser microdissected regions were comparable in size the amount of isolated DNA varied between 1.8 ng for grey and light blue areas and 6 ng for blue and dark blue areas. Copy analysis was done by TaqMan qPCR with 250 pg laser microdissected DNA for each experiment. Analyses were performed using three biological replicates, each with two technical replicates, and undifferentiated hMSC DNA served as the control. Copy number variations for AGAP2, CDK4, CPT1B, TRIB1, and SHANK3 are shown in Fig. 4E, and corresponding calculated P-values were summarized in Tables I and II. A significantly increased CDK4 copy number was observed in cells from grey and light blue regions, corresponding to absent or low proteoglycan content. CPT1B copy number was significantly elevated in blue regions, indicative of proteoglycan deposition. In addition, an increase in SHANK3 copy number was detected in blue regions, reflecting higher proteoglycan levels. Interestingly, dark blue areas with highest proteoglycan levels revealed no significant copy number increase for the investigated genes. AGAP2 showed no copy number increase in the laser microdissected probes for all staining intensities. TaqMan analysis of MDM2 revealed no results for technical reasons. Unfortunately, the very low amount of DNA from microdissected areas is a limitation of this study since chondrocyte marker analysis was not possible.

Table I One-way ANOVA results for AGAP2, CDK4, CPT1B, SHANK3 and TRIB1 copy number values grouped by laser-microdissected areas with different toluidine blue staining intensities.

Table I

One-way ANOVA results for AGAP2, CDK4, CPT1B, SHANK3 and TRIB1 copy number values grouped by laser-microdissected areas with different toluidine blue staining intensities.

Gene	Term	Degrees of freedom	Sum of squares	Mean of squares	F-value	Pr (>F)
AGAP2	Staining	4	0.2175	0.0544	0.3395	0.8487
	Residuals	25	4.0042	0.1602
CDK4	Staining	4	11.6078	2.9019	6.6149	0.0011a
	Residuals	23	10.0901	0.4387
TRIB1	Staining	4	1.3868	0.3467	1.5625	0.2164
	Residuals	24	5.3256	0.2219
SHANK3	Staining	4	11.0040	2.7510	2.7278	0.0689
	Residuals	15	15.1273	1.0085
CPT1B	Staining	4	7.3789	1.8447	3.0215	0.0409a
	Residuals	21	12.8213	0.6105

[i] ^aP<0.05. A significant P-value indicated that the means between the groups were different; further analysis was performed using the Tukey HSD post-hoc test (Table II).

Table II Tukey's HSD post hoc test for staining pairs with significant P-values (adj P<0.05).

Table II

Tukey's HSD post hoc test for staining pairs with significant P-values (adj P<0.05).

Gene	Comparison	Mean difference	CI lower 95	CI upper 95	Adj. P-value
CDK4	Light blue vs. CTRL	1.6367	0.5063	2.7671	0.0024
CDK4	Grey vs. CTRL	1.4867	0.2228	2.7505	0.0158
CPT1B	Blue vs. CTRL	1.5217	0.1778	2.8656	0.0215
CDK4	Dark blue vs. light blue	-1.2700	-2.4004	-0.1396	0.0225
CDK4	Blue vs. light blue	-1.1650	-2.2954	-0.0346	0.0412

Discussion

Gene amplification has long been associated with tumorigenesis and multidrug resistance, although its physiological role was first recognized in Drosophila melanogaster during eggshell development (20,21). Here, we identified gene amplifications in hMSCs during chondrogenic differentiation. Previous fluorescence-in-situ-hybridization (FISH) analyses of hMSCs differentiating into adipocytes and osteoblasts revealed heterogenous populations, with amplified and non-amplified cells coexisting (2). However, FISH analysis proved challenging for chondrocyte differentiation due to the three-dimensional growth pattern and proteoglycan production, which interfere with FISH. Here, we investigated gene amplification during the first seven days of chondrocyte differentiation using qPCR and Taq Man assays targeting genes previously shown to be amplified during osteoblast and adipocyte differentiation, as well as genes that were known to be amplified in chondrosarcoma (2,6).

DNA sequence amplification, as defined by Bostock CJ in 1986, refers to any event that increases a gene's copy number per haploid genome above the level characteristic for that organism (1). In accordance with a criterion defined in our recent study (22), qPCR-derived copy number values exceeding 2.3 were considered indicative of gene amplification. To distinguish gene amplification from whole-chromosome duplication, we included genes located on the same chromosomes (AGAP2, CDK4, and MDM2 on chromosome 12; CPT1B and SHANK3 on chromosome 22; and TRIB1 and MYC on chromosome 8).

During chondrocyte differentiation, qPCR analysis revealed amplification of several investigated genes at days 2 and 5, followed by a decline by day 7. To assess stage-specific amplification, we performed laser microdissection of toluidine blue-stained areas after five days of differentiation, when proteoglycan deposition was evident. Cells were grown on PET membrane slides, enabling precise isolation of regions with varying staining intensities.

Amplification patterns differed across differentiation stages: CDK4 showed the highest copy number in regions lacking and low proteoglycans, indicating early activation, whereas SHANK3 peaked in areas with strong proteoglycan staining, suggesting stage-specific amplification. Regions with the most intense staining displayed no significant amplifications for all tested genes, consistent with cessation of amplification during terminal differentiation. Although AGAP2 increased in bulk DNA at days 5 and 7, it was undetectable in microdissected samples, and MDM2 yielded no results, likely due to technical limitations of microdissection and DNA isolation. A possible reason for failure of MDM2 qPCR amplification could be the very low amount of DNA per reaction (250 pg) and degradation effects during laser microdissection and DNA isolation. In this respect we would like to mention that gene amplification analysis using microdissected samples is a starting point to address gene amplifications in a heterogenous cell population but has limitations because of low DNA yield and degradation during laser exposure. Future studies need to be expanded to increase DNA yield and to have the possibility to test more genes on gene amplification including chondrocyte marker genes and additional genome regions. Furthermore, although CPT1B and SHANK3 are both located on chromosome 22, their distinct amplification patterns across microdissected regions argues against whole-chromosome duplication and instead support gene-specific amplification events.

Chondrosarcoma development is closely linked to the dysregulation of normal chondrogenic differentiation pathways. Chondrosarcoma is increasingly viewed as a disease in which molecular alterations accumulate along the differentiation trajectory of chondrogenic cells rather than arising independently of cartilage development. During early differentiation mutations alter proliferation and differentiation of chondrocytes and very likely support malignant transformation. Those early mutations were supposed to alter the IHH/PTHrP and IDH1/IDH2 pathways (23). Late-stage mutations were suggested to involve cell cycle regulating pathways including p53 and RB (23). As early as 1980 Kreicbergs et al (24) found that many highly malignant and metastasizing chondrosarcoma were hyperploid. In this respect it is interesting that we found CDK4 amplifications predominantly during early differentiation stages and no significant gene amplifications during the latest differentiation stage. So early differentiation associated mutations can lead to unscheduled CDK4 amplification with known growth advantages whereas late differentiation mtations especially TP53 mutations probably can disturb proper cease of gene amplification and lead to genome instability.

In addition, for normal chondrocyte differentiation of mesenchymal stem cells, a precise timing of induction and silencing of SOX9 expression is necessary (25). Here we found an increasing SOX9 expression during differentiation with highest SOX9 expression after 7 days. Posttranscriptional regulation of SOX9 expression through altered miR-145 expression was described as epigenetic regulation of SOX9 in human chondrosarcoma (26). We can speculate if this regulation occurs late during differentiation maybe a loss of SOX9 gene function thereby leads to additional genome instability, polyploidy and further promotes tumorigenesis.

Despite tumor development chondrocyte differentiation may also affect Osteoarthritis (OA) and extracellular matrix (ECM) dysregulation. OA is mainly a disease of extracellular matrix (ECM) imbalance (27). Therefore, any gene affecting chondrocyte proliferation and metabolism of ECM proteins, could potentially contribute to disease susceptibility or progression.

Chondrocytes were the main producer of ECM proteins. During OA progression chondrocytes secrete inflammatory cytokines that were involved in disruption of cartilage homeostasis and cartilage degradation (28). Several studies report an association between OA and fibrosis (29). During normal chondrocyte differentiation CDK4 amplification likely necessary for regulating the balance between proliferation and differentiation. Since CDK4 gene amplification is no more detectable in late differentiation stages and as mentioned above for chondrosarcoma development mutations can lead to growth advantage of cells with unscheduled CDK4 amplification resulting in proliferation of chondrocytes and imbalances of ECM proteins and inflammatory cytokines.

CPT1B is a rate-limiting enzyme for mitochondrial fatty-acid β-oxidation and was recently shown to be highly expressed in myoblasts of OA patients (30). During normal differentiation CPT1B amplification is probably necessary for an enhanced metabolism and for OA development the involvement is still unknown. SHANK3 is known as a postsynaptic scaffold protein in neurons and an involvement in normal cartilage development is unknown. We can only speculate that during normal chondrocyte differentiation SHANK3 is necessary for cytoskeletal organization or cell-matrix communication but a functional role in OA development remains unclear.

In conclusion, we observed in both bulk and microdissected analyses that gene amplification is an early, transient process during chondrogenic differentiation, diminishing as cells mature-likely to preserve genomic stability. Since this study has limitations, further orthogonal validation methods including single cell analysis were needed. Future studies should explore mechanisms that regulate amplification cessation and their implications for regenerative therapies. Nevertheless, the current work provides valuable insights into advantages and capabilities including risks of these cells.

Acknowledgements

Not applicable.

Funding

Funding: This work was supported by the Universität des Saarlandes (grant no. 2025-Fischer) to UF and the Deutsche Forschungsgemeinschaft (DFG; grant no. 447452855 INST 256/541-1) to EM.

Availability of data and materials

The data generated in the present study may be found in the SRA database under BioProject accession number PRJNA1504250 or at the following URL: https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA1504250.

Authors' contributions

PES performed the investigation, and reviewed and edited the manuscript; MC performed the investigation, and reviewed and edited the manuscript; SR performed formal analysis, including statistical analysis and sequencing analysis, and reviewed and edited the manuscript; AK was involved in conceptualization, acquired funding, and reviewed and edited the manuscript; EM was involved in conceptualization, acquired funding, wrote the original draft, and reviewed and edited the manuscript; UF was involved in conceptualization, performed the investigation, wrote the original draft, and reviewed and edited the manuscript. SR and AK confirm the authenticity of all the raw data. All authors read and approved the final manuscript.

Ethics approval and consent to participate

The in vitro use of primary human cells from PromoCell does not require ethics approval from the institutional ethics committee; this does not violate the ethical, legal or political standards or principles within Germany. PromoCell's donor acquisition, documents and cell collection process has been assessed and evaluated by the ethical commission of the doctor's medical association of Baden-Württemberg (ethic approval nos. # 219-04 and # F-2020-157).

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

References