This week’s journal club focuses on a study from Stem Cell Reports (2026) aiming to establish a foundation for harmonized quality control for induced pluripotent stem cell (iPSC).
by Agnes Sandvik
Aim:
The authors point out that quality control of iPSCs is inconsistently evaluated across different laboratories. Thereby the aim of their work was to build an empirical foundation to standardize iPSC quality control. They wanted to benchmark how reproducible results were across different labs, pinpoint the most reliable pluripotency markers and generate data-driven guidance for quality control so that all may use a common practice to make it possible to create iPSC banks that can be utilized across labs and countries without issues to ensure reliable and accessible supply. They also think this could help to make the laws align between different governments and help to upscale the production of treatment/therapies.
Methods:
To address this, they conducted two international Quality Assessments to evaluate qPCR-based genomic integrity testing and flow cytometry-based assessment of the undifferentiated state.
Results:
Current testing methods for iPSC lines (how good and standardized are the tests? Does the different laboratories conduct the same tests?)
QAR 2019:
Out of 18 participants only 5 (28 %) of the participants used formal national or international quality standards for iPSC. Genetic integrity testing (karyotype, copy number variation, DNA methylation profiling and whole-genome sequencing) was performed by 14/18 (78 %), but only 6/14 (43 %) reported using a standard operating procedure (SOP). Phenotype characterization (through immunocytochemistry and flow cytometry) was performed by 14/18 (78 %) but only 7/14 (50 %) used SOP for at least one of the methods.
Collectively this indicates that while quality assessment is done by a majority of the participants, a need to standardize and validate testing protocols to ensure consistency and reliability of iPSC characterization across sites remains.
QAR 2023: (What’s new?)
Only 4/18 (22 %) of the participants used formal national or international quality standards for iPSC (slight decrease). Genetic integrity testing was performed by 13/18 (72 %) (also a bit lower) but 9/13 (69 %) reported having SOP in place (a bit better). Phenotypic characterization was performed by 12/18 (67 %) (less than 2019) but 11/12 (92 %) use SOP (huge improvement). Now genotyping is also performed by 12/18 (67 %) and 8/12 (67 %) use SOP for this.
This indicates that iPSC testing practices still vary widely between participants. This lack of consistency highlights the need for standardized protocols and clear regulatory frameworks to ensure the safety, reproducibility, and scalability of future iPSC banks.
Genetic integrity testing with a commercially available kit (how well will different participants discover genetic faults in iPSC by using a commercial qPCR-kit?)
In QAR 2019 each participant tested two blinded genomic DNA samples using a commercially available qPCR kit, designed for detection of 8 recurrent genomic abnormalities in hPSC cultures.
- CiRA G1 (normal diploid karyotype)
- CiRA G2 (unbalanced structural rearrangement resulting in gain of the long arm of chromosome 1 (1q))
In total 45 runs were completed as the different participants chose to complete different amounts of runs.
Fig 1 gives a collected overview of the median copy number across each region. 42/45 (93 %) correctly identified the amplification of chromosome 1q in the CiRA G2 sample, while accurately classifying the CiRA G1 sample as diploid. Indicating that the test functions quite well. Statistical analysis showed no significant difference in detected copy number between the samples, apart from 1q that was much higher in G2 (p < 0.0001) as expected.
iPSC characterization by flow cytometry (how different participants analyze iPSC)
QAR 2019:
18 labs got the same cell samples:
- CiRA F1 (clean, undifferentiated cells)
- CiRa F2 (cells that had started differentiating)
- NIBSC (technical copies, same sample repeated several times)
and used their own methods to analyze them. The goal was to see how similar their results would be.
The authors set the requirement that at least 4000 cells had to be measured in flow cytometry, 60 % had to be in the right area (P1 gate). Only 4 labs managed this for NIBSC samples, the rest were excluded from further analysis.
Statistics:
A mixed-effects model was used to estimate differences between CiRA F1 and NIBSC samples, taking into account variability across markers and participants. NIBSC samples were almost identical (0.1 % difference), but had slightly lower stemcell signal than CiRA F1 (-1.5 %)
When testing the undifferentiated samples most labs got the same results. In contrast, assessment of the spontaneously differentiated CiRA F2 sample exhibited substantial variability between the participants. Among the utilized markers, NANOG had the greatest inter-laboratory variability, while OCT3/4 and TRA-1-60 showed the least.
The most commonly used markers were OCT3/4, TRA-1-60 and SSEA4, followed by TRA-1-81, SOX2, NANOG and SSEA3.
Main problem in 2019: flow results were very inconsistent between different labs. Therefore, they did QAR 2023: testing if standardized methods yield more similar results.
23 labs participated, 3 cell samples (undifferentiated, differentiated, 1:1 mix) were analyzed by all participants using two separate quality tests:
- Using their own methods
- Same standardized method for all participants
Quality Test 1- using their own methods:
Some participants used 3 while others used 4 markers:
- Everyone used OCT3/4
- 86 % used TRA-1-60
- 82 % used SSEA4
- 41 % used SOX2.
There was variation in testing the samples 1, 2 or 3 times. 15 different flow cytometry machines were utilized across participants.
The percentage of positive cells for each marker tested was calculated as the mean across all participants’ data (Fig 2). As the mixed sample combined undifferentiated and differentiated cells in a 1:1 ratio it was expected that that the marker expression would vary proportionally to this. CD45, SOX2, SSEA4 and TRA-2-49 showed low variability between different participants, but their expression levels remained consistent between all the samples (differentiated, undifferentiated and mix) proving them to be not very useful in distinguishing between differentiated and undifferentiated cells.
Conversely NANOG and TRA-1-81 showed proportional changes in expression correlating with differentiation, but these markers displayed higher inter-participant variability. Notably four markers CD30, OCT3/4, SSEA5 and TRA-1-60 combined low inter participant variability with clear proportional changes in expression across samples. Indicating them suitable for assessing differentiation status.
To investigate further which markers most effectively distinguish differentiated from undifferentiated stem cells, the percentage of marker-positive cells in the differentiated samples were subtracted from that in the undifferentiated samples. Yielding the difference in marker expression (Table 1). The higher the difference, the better the marker is at differentiating the samples from each other.
They set the threshold to markers that exhibited a difference in mean greater than 30 % as suitable for confirming stem cell differentiation. Markers exceeding this threshold include SSEA5, TRA-1-60, TRA-1-81, OCT3/4, CD 30 and NANOG.
To evaluate marker consistency, markers showing no change across samples or analyzed by only one participant was excluded. They were left with markers: NANOG, OCT3/4, TRA-1-60 and TRA-1-81 which were analyzed to see how consistent the measurements were within each sample. Highest variation between participants for differentiated cells were seen for NANOG, OCT3/4 and TRA-1-60 while for undifferentiated cells there was higher variability in TRA-1-81
In this quality test, even though all participants used the same cells they got different results as the methods were not standardized.
Quality Test 2- same standardized method for all participants:
All 23 laboratories used the same chemicals, fluorophores, staining procedure and analysis method and all but one of the participants (“Swallow”) classified the samples correctly.
To evaluate the effectiveness of the four markers TRA-1-60, SSEA4, OCT3/4 and PAX6 they estimated how well the 1:1 differentiated and undifferentiated samples correlated to the theoretically expected result. They found that the measurements of the mixed sample were exactly as theoretically expected (Fig 3).
89.7 % of results were within an accepted range 80-125 % of the expected values. Indicating that the markers performed consistently and proportionally between the participants.
As previously observed, one lab (“Swallow”) had completely different results. Further investigation revealed that shipment of samples to this lab group took double the time as for the others, this may have compromised reagent integrity and the quality of the cell samples and could explain the faulty results.
Mean percentage of cells positive for each of the four markers TRA-1-60, SSEA4, OCT3/4 and PAX6 was calculated (Fig 4), excluding data from the “Swallow” lab group . The most variable marker was PAX6, followed by TRA-1-60, OCT3/4 and SSEA4.
Fig 4: most participants’ data show strong agreement across markers.
SSEA4: The marker SSEA4 was used by 18/22 participants in Quality Test 1 and was also included in Quality test 2. Notably, the expression level remained high across all analyzed samples. As it is one of the markers commonly recommended for inclusion in iPSC QC panels based on expert consensus and given its widespread use among participants the authors conclude that it should be further investigated how relevant it really is to use this marker. Therefore, a validation run was performed to examine marker expression in greater detail.
Validation round 2024
Building on findings from QAR 2023 the expression of SSEA4, OCT3/4, TRA-1-60 and SSEA5 was assessed in five different iPSC lines differentiated into three germ layers (Fig 5).
SSEA4 remained high for all lines at day 0, day 5 endoderm, day 5 mesoderm and day 7 ectoderm. Consistent with previous findings OCT3/4 and TRA-1-60 expression decreased rapidly in comparison to SSEA4 with levels decreasing from day 5 in all cell lines. SSEA5 also decreased from day 5 with 0% expression observed in day 5 mesoderm and day 7 ectoderm samples. Day 5 endoderm had slightly higher expressions of OCT3/4, TRA-1-60 and SSEA5 than in other lineages, likely because of lower efficiency In endoderm differentiation, yet still markedly lower than SSEA4.
To assess if SSEA4 expression decreased over time they did several longer differentiations and found a progressive decrease in expression across most cell types over time (Fig 5). Expression of TRA-1-60, OCT3/4 and SSEA5 remained low across all long-term differentiation conditions.
Conclusions
- Global iPSC quality control standards are still fragmented, limiting scalable manufacturing of iPSC-derived advanced therapy medicinal products (ATMPs).
- Multi-site benchmarking provides the first empirical evidence that both flow cytometry and qPCR can be reproducible when protocols are standardized.
- OCT3/4, TRA-1-60 and SSEA5 proved to be the most reliable pluripotency markers, with the lowest inter-lab variability and clear proportional changes between differentiated, undifferentiated and mixed samples.
- SSEA4 remains highly expressed in all cell states, making it useful for detecting residual undifferentiated cells in finished products, but less effective for monitoring early differentiation.
- Recommended quality control of iPSC include; G-band karyotyping with genome-wide molecular assays and qPCR as a routine screen.
- International collaboration is essential to turn these data-driven findings into accepted standards and to create international iPSC banks.
Continue your reading here:
Toward standardized iPSC testing: Insights from a multi-year international Quality Assessment Round
Hägg A, Wood R, Mochizuki AL, Abberton K, Abranches E, Alvarez-Palomo B, Baptista R, Quintanilha Barbosa RA, Barry J, Carvalho AB, Bennaceur Griscelli A, Campos de Carvalho AC, Chaker D, Chang H, Choi HY, Codinach M, Arán Corbella B, Cowan S, Dickerson SJ, Elwood N, Fan X, Feyeux M, Forrester M, Gaffney A, Guilbert SM, Ha HY, Hirst AJ, Hunter AL, Jamieson LG, Judson RN, Kanemura Y, Kasai-Brunswick TH, Kim JH, Kim H, Kintali M, Krishnan S, Kuebler B, Lau CY, Li W, Mack A, MacLeod MR, Madrid M, Mamiya H, Manache-Alberici L, Mărginean D, Mentre O, Morgan SL, Mountford J, Munir H, Ng SHS, Ogawa H, Oh S, Ohara H, Oono K, Park N, Pereira LV, Pereira da Silva Bezerra I, Podovei AR, Querol S, Raje J, Raya A, Sakamoto S, Sarafian R, Schmit K, Selvitella S, Singh G, Smart MJK, Song J, Stacey G, Sullivan S, Sumida M, Terrenoire C, Tian P, Uhlin E, Vaquero JMA, Veiga A, Vicky Wang JW, Warre-Cornish K, Wood J, Yamamoto A, Zhang G, Hikichi T, Turner M, Falk A.
Stem Cell Reports. 2026 Apr 14;21(4):102857. doi: 10.1016/j.stemcr.2026.102857. Epub 2026 Mar 19.PMID: 41861820