Abstract
Cultured meat presents a sustainable alternative to traditional meat
production but faces significant challenges in scalability and cost
efficiency. A key limitation is the restricted proliferation capacity
of bovine mesenchymal stem cells (bMSCs), a widely used cell source in
the field. Using a pooled, lentiviral CRISPR knockout screen, we
interrogated 3000 CRISPR guides targeting 600 genes involved in stem
cell regulation or proliferation. Notably, knockouts of TP53 and PTEN
significantly increased proliferation rates and delayed senescence.
Validation with individual gene knockouts confirms these effects,
showing enhanced growth but reduced differentiation potential. We also
identified chondrogenic differentiation as a promising target whose
repression may further promote MSC expansion. These findings
demonstrate the utility of CRISPR screening for optimizing bovine stem
cell traits and offer a path toward more scalable cultured meat
production in the future.
Subject terms: Stem-cell biotechnology, Mesenchymal stem cells,
High-throughput screening
__________________________________________________________________
Pooled CRISPR knockout screening in bovine mesenchymal stem cells
identifies TP53 and PTEN as key regulators of proliferation, offering a
genetic strategy to enhance cell expansion and address scalability
challenges in cultured meat production.
Introduction
The global demand for meat continues to rise^[36]1, with projections
suggesting a 14% increase in production and consumption from 2020 to
2030^[37]2,[38]3. However, conventional animal agriculture presents
significant environmental and public health challenges^[39]4, including
intensive resource utilization (land, water, and energy), driving
deforestation and biodiversity loss^[40]5–[41]7, substantial greenhouse
gas emissions, contribution to antibiotic resistance, and potential for
zoonotic disease transmission^[42]8. In response to these challenges,
cultured meat, also known as lab-grown or cultivated meat, has emerged
as a promising alternative^[43]9,[44]10. This innovative approach aims
to provide a sustainable and humane substitute for traditional meat
production by in vitro proliferation of animal cells in a controlled
environment, eliminating the need for live animals. Cultured meat
relies on stem cells, which can both proliferate exponentially and
differentiate into specialized, edible tissues such as muscle
(myocytes) and fat (adipocytes)^[45]11,[46]12. The process involves
extracting stem cells, adapting them to in vitro conditions, expanding
their biomass, and inducing specialization into mature tissues for
final processing. While this approach offers potential solutions to the
issues associated with conventional meat production^[47]13,[48]14, the
current state of cultivated meat faces significant challenges,
including high costs, scalability issues, and the need for further
technological advancements^[49]15.
This study addresses one of the key challenges in cultured meat
production: developing efficient, scalable cell lines^[50]16. Various
cell types are utilized for this purpose, including mesenchymal stem
cells (MSCs)^[51]17, embryonic stem cells (ESCs), induced pluripotent
stem cells (iPSCs), and primary muscle or adipose cells. ESCs and iPSCs
offer broad differentiation potential but face challenges related to
complexity, cost and differentiation efficiency. Primary differentiated
cells exhibit efficient lineage commitment but are limited in
proliferation capacity and scalability. MSCs offer a practical balance
of multipotency, ease of culture, and industrial relevance, accounting
for approximately 25% of cell sources used in the cultured meat
sector^[52]18. They are readily isolated from adult tissues^[53]19, and
can differentiate into mesodermal lineages such as osteocytes (bones),
chondrocytes (cartilage), adipocytes (fat), or myoblasts
(muscle)^[54]20, with well-established protocols for culture and
differentiation^[55]21. However, their limited proliferation capacity
remains a bottleneck for large-scale application.
MSCs face two key proliferation constraints: slow intrinsic growth
rates and replicative senescence. While MSCs divide more rapidly
compared to terminally differentiated primary cells, their doubling
times remain substantially longer than those of pluripotent stem
cells^[56]16,[57]22, implying underutilized molecular mechanisms could
accelerate their division. A second limitation arises from progressive
telomere shortening and stress accumulation^[58]23, which drive
irreversible senescence, a growth arrest state restricting large-scale
expansion^[59]24. To address these challenges, strategies must either
enhance baseline proliferation through cell-cycle manipulation or delay
senescence by mitigating aging mechanisms. Targeted genetic editing of
these dual processes could optimize MSC expansion efficiency while
maintaining their essential multipotency and differentiation potential,
ultimately advancing scalable cultured meat systems.
To optimize MSC proliferation, we employed a CRISPR knockout screening
approach, which enables systematic perturbation of thousands of genes
to identify those with significant physiological roles^[60]25,[61]26.
In pooled screens, a complex library of guide RNAs and Cas9 is
delivered to a heterogeneous cell population, allowing cells with
advantageous mutations to be enriched through competition. In contrast,
arrayed screens test each sgRNA in separate wells, enabling precise
genotype-to-phenotype mapping, but at significantly higher cost and
lower throughput^[62]23. Another key consideration is the scope of the
screen: whole-genome versus targeted. While whole-genome screens
provide broad coverage across the transcriptome, they are more
difficult to scale and interpret, as the majority of sgRNAs have no
functional effect. We selected a pooled, targeted approach to balance
scalability, cost-effectiveness, and the ability to leverage
proliferation as a straightforward enrichment phenotype. By enabling
cells to compete in a shared environment, this strategy efficiently
uncovers genetic factors driving MSC proliferation, potentially
advancing scalable cultured meat production.
Our study aims to expand the application of CRISPR screening in bovine
cells, focusing specifically on enhancing proliferation in bMSC, and
thus addresses one of the key challenges in cultured meat production:
scalability and cost-effectiveness. This approach has the potential to
significantly improve the efficiency of cultivated meat, bringing us
closer to a sustainable alternative to traditional meat production that
can meet the growing global demand for protein while mitigating
environmental and health concerns associated with conventional
livestock farming.
Results
Building a bovine-specific CRISPR screening system: MSC isolation,
validation, transduction, and CRISPR activity assessment
To enhance the proliferation rate of bMSCs for cultured meat
production, we performed a CRISPR knockout screen to identify key genes
and pathways regulating cell proliferation (Fig. [63]1A). This screen
required both isolated MSCs and a pooled CRISPR single-guide RNA
(sgRNA) library targeting 603 genes, including transcription factors
and proliferation-related genes. In this study, we specifically focus
on adipose-derived bMSCs, as their enhanced adipogenic differentiation
potential makes them particularly suitable for generating fat tissue, a
critical component of cultured meat.
Fig. 1. Experimental design and characterization of bovine MSCs.
[64]Fig. 1
[65]Open in a new tab
A Experimental design overview of CRISPR-based screening in bovine
MSCs. B Growth kinetics of isolated bMSCs showing consistent cell
doubling rate up to 100 days in culture, followed by population
doubling arrest, measured by cell counting (n = 3). Blue line and
shaded envelope represent the fitted linear model and its 95%
confidence interval, respectively. C Boxplot showing the percentage of
senescence-positive cells in early-passage, late-passage, and
etoposide-treated early-passage cells. Senescent cells were
significantly increased in both late-passage and 20 µM
etoposide-treated groups compared to early-passage controls. Each point
represents one image field from two biological replicates (n = 2). D
Multi-lineage differentiation potential of bMSCs. Left: Chondrogenesis
induced by 21 days in chondrogenic media, visualized by alcian-blue
staining of proteoglycans. Right: Adipogenesis induced by 21 days in
adipogenic media, visualized by oil-red O staining of lipid droplets. E
Flow cytometry analysis of bMSCs surface markers CD29 (APC-labeled
primary antibody) and CD44 (APC-labeled secondary antibody) at passage
6 (n = 3 for each). >90% of the population expresses both markers. Gray
plots represent cells stained with secondary antibody only. F GFP
expression in bMSCs and HEK293T cells following lentiviral
transduction. Green histograms represent GFP-transduced cells (n = 3),
gray histograms show non-transduced control cells (n = 2). G Cas9
expression in transduced bMSCs assessed by mRNA reverse
transcription-PCR, controlled by GAPDH and RPS2 housekeeping genes
expression (n = 2).
Accordingly, we first isolated and characterized adipose-derived bMSCs
(AD-bMSCs) from a 3-month-old calf, following an established
protocol^[66]27. To evaluate the proliferative capacity of the isolated
bMSCs, we tracked population doubling time (PDT) over 156 days.
Proliferation remained stable until day 79, with an average PDT of
37.7 ± 8.0 h. After day 79, the doubling rate progressively declined,
indicating a gradual loss of proliferative capacity (Fig. [67]1B). To
determine whether this decline was connected to senescence, we
performed senescence-associated β-galactosidase (SA-β-gal) staining at
early (day 20) and late (day 80) passages. A significant increase in
the percentage of senescent cells was observed in late-passage compared
to early passage cells (average of 35% and 11.7% respectively), while
etoposide-treated control cells showed a high percentage of senescent
cells (87.8%), indicating that at least some of the decline in
proliferation rate can be attributed to an increase in population
senescence (Fig. [68]1C).
We next tested the cells’ stemness by verifying their differentiation
potential into adipocytes and chondrocytes (Fig. [69]1D) and by
expression of MSC surface markers CD29 and CD44 (Fig. [70]1E,
p < 0.001, t-test). To assess the genetic engineering capabilities of
AD-bMSCs, we evaluated their transduction using lentivirus. Two
plasmids were tested: one expressing GFP and the other expressing Cas9
nuclease. Flow cytometry confirmed successful GFP transduction in both
bMSCs and HEK293T cell lines, with approximately 49% of bMSCs and 80%
of HEK293T cells showing significantly higher green fluorescence
compared to untransduced populations (p = 0.004 and p < 0.001,
respectively, t-test, Fig. [71]1F). Related data, including gating
strategy, senescence imaging, additional adipogenesis analysis, and
transduction optimization are shown in Fig. S[72]1. Cas9 expression was
verified by reverse transcription PCR of RNA from Cas9-transduced cells
(Fig. [73]1G). With successful transduction and Cas9 expression
confirmed, we proceeded to construct a targeted sgRNA library to
identify genes involved in MSC proliferation.
CRISPR short-term proliferation screen enriches and depletes known and novel
genes
We designed a focused bovine-specific CRISPR knockout library to
enhance MSC proliferation by targeting 603 Bos taurus genes related to
proliferation (Fig. [74]2A). The library included transcription
factors^[75]28 and genes from analysis of the Dependency Map (DepMap)
portal^[76]29 ([77]https://depmap.org/portal) (Fig. S[78]2A, B).
DepMap-derived genes were chosen by ranking their mean Gene Effect
Score (GES), akin to fitness, across all cell lines in DepMap and
selecting the highest 100 (“Top Gene Effect Score/TGES”) and lowest 50
(“Bottom Gene Effect Score/BGES”) CRISPR score-ranked genes, which are
predicted to increase or decrease proliferation upon knockout,
respectively. The mean Gene Effect Score from the DepMap database is
shown by group (Fig. [79]2B), illustrating the predicted impact of gene
knockout on cell viability for each category. The final library
comprised 2564 sgRNAs targeting 603 coding regions (3–5 sgRNAs per
gene), with an additional 436 control sequences targeting non-exonic
regions. sgRNA sequences were retrieved and filtered from a
whole-genome CRISPR library^[80]30, then NGS and barcoding adapters
were designed and added to the sequences. Subsequently, oligos were
synthesized, amplified, and cloned into a lentiviral vector (Fig.
S[81]2C, D). Finally, in preparation for the CRISPR screen experiment,
library lentivirus particles were produced for MSC transduction.
Fig. 2. CRISPR library design and screening in AD-bMSCs reveals TP53 and PTEN
as key regulators of cell proliferation.
[82]Fig. 2
[83]Open in a new tab
A Composition of the designed CRISPR sgRNA library. The library
consists of 3000 sgRNAs targeting 603 genes, grouped into:
Transcription Factors (TF, 2099 sgRNAs/484 genes), Top Gene Effect
Score (TGES, 304 sgRNAs/73 genes), Bottom Gene Effect Score (BGES, 161
sgRNAs/46 genes), and controls (436 sgRNAs). B Mean Gene Effect Score
across all DepMap datasets for each library subgroup, indicating
predicted impact on cell fitness upon gene knockout. C, D CRISPR screen
enrichment analysis. AD-bMSCs were transduced with Cas9-expressing
vector followed by the designed sgRNA library (n = 3). AD-bMSCs were
sampled on days 3, 7, 10, 16, 24, and 30 post-library transduction. C
Minimum hypergeometric test (mHG) enrichment analysis. Log2 fold-change
(y-axis) plotted against log2FC rank (x-axis) over time. Genes with top
Gene Effect Scores (TGES) showed significant enrichment (red), while
genes with bottom scores (BGES) showed significant depletion (blue).
mHG p values are displayed for each day: TGES enrichment (top left,
red) and BGES depletion (top right, blue). D Volcano plots of screen
results by day. −log10 p value (x-axis) plotted against log2
fold-change (y-axis), calculated using MAGeCK. Red: significantly
enriched gene-knockouts (leading with TP53 and PTEN); Blue:
significantly depleted gene-knockouts; Gray: non-changing
gene-knockouts. Dashed lines indicate significance thresholds
(p < 0.05, |log2FC | > 1).
Using Cas9-expressing bMSCs and a lentiviral sgRNA library, we
conducted a CRISPR screen to identify genetic factors influencing cell
proliferation. The bMSCs were transduced with the sgRNA library at a
low multiplicity of infection (MOI ~ 0.3), ensuring coverage of
approximately 600 cells per sgRNA (Fig. S[84]2E). Over a 30-day
screening period, cells were passaged every 3–4 days, with DNA samples
collected at specified intervals for sgRNA quantification via
next-generation sequencing (NGS). sgRNA quantification, ranking, and
statistical testing were performed using MAGeCK^[85]31.
A minimum hypergeometric^[86]32 (mHG) analysis of the predefined
library sub-groups showed significant enrichment or depletion patterns
(p < 0.001 | *** | for most days). Gene knockouts enhancing
proliferation in the DepMap database, termed TGES, were found to be
enriched in the screen, and vice versa, those decreasing proliferation
(termed BGES) were depleted (Fig. [87]2C). Generally, genes that
enhance proliferation in human cancer cells (TGES) were similarly
enriched in the bovine MSC screen, while genes that reduce
proliferation (BGES) were correspondingly depleted, demonstrating
consistent functional roles across cell types. Transcription factors
showed varied responses: some enhanced proliferation while others were
quickly depleted.
To evaluate gene-specific effects on survival and proliferation, we
used MaGeCK to calculate fold changes based on the aggregated
performance of each gene’s 3–5 sgRNAs relative to baseline (day 0;
Fig. [88]2D). This analysis revealed significant shifts in sgRNA
abundance over time, identifying candidate proliferation inhibitors
such as TP53, PTEN, and VGLL4 (p < 0.001 at later time points), whose
knockouts increased in abundance, suggesting a growth advantage. In
contrast, sgRNAs targeting essential genes for proliferation, including
ribosomal proteins and JUN, decreased over time (p < 0.001), consistent
with reduced cell fitness. Among all candidates, TP53 knockout produced
the most pronounced effect, with an over 1000-fold increase in
abundance by day 30. This is consistent with TP53 established role as a
tumor suppressor that induces cell cycle arrest and senescence in
response to stress^[89]33. PTEN, similarly, acts as a negative
regulator of the PI3K/AKT pathway, where its loss leads to increased
mTOR activity and can trigger a senescence-like response in some
contexts^[90]34. The enrichment of sgRNAs targeting TP53 and PTEN in
our screen supports their well-characterized roles in restraining cell
growth and highlights their importance in maintaining proliferative
homeostasis in MSCs.
Pathway analysis of short-term CRISPR screen identifies five clusters and
associated pathways
Building on the insights gained from our initial CRISPR screen, which
identified key singular genetic factors influencing cell proliferation,
we sought to deepen our understanding by conducting a gene pathway
enrichment analysis to identify the regulatory pathways governing the
proliferation of AD-bMSCs. This analysis identified five distinct gene
clusters based on Spearman’s correlation of change over time (Fig.
S[91]3A, B). Each cluster exhibited unique abundance patterns over time
(Fig. [92]3A, B). Notably, Cluster 1, which consists of 165 genes with
monotonically increasing abundance, was of particular interest for
potential enhanced proliferation. Other significant clusters included
Cluster 3, characterized by a decrease in gene abundance, and Cluster
2, which initially decreased followed by an increase. We further
characterized these clusters using a ranked over-representation assay
(ORA) with the gprofiler2^[93]35 tool to identify enriched Gene
Ontology Biological Process (GO:BP) terms (Figs. [94]3C and S[95]3C).
The gene ranking for ORA was determined based on the maximal p value
for each gene (Fig. [96]3C, down). Cluster 1 was significantly enriched
with nine GO:BP terms related to proliferation, including three terms
associated with TGFβ production and chondrocyte differentiation,
processes which are relevant to MSC function and differentiation. These
findings suggest potential pathways and gene targets for manipulation
to enhance MSC proliferation. However, these findings apply to
short-term culturing, and we were also interested in exploring the
effects of long-term culturing, particularly through the application of
a second CRISPR screen.
Fig. 3. Temporal gene expression patterns and functional analysis of CRISPR
screen results.
[97]Fig. 3
[98]Open in a new tab
A Heatmap showing log2 fold-change trajectories over 30 days for all
603 genes. Red colors represent enriched genes, and blue colors
represent depleted genes, with color intensity indicating magnitude of
change. Clusters (1–5) correspond to those in (B). B Mean log2
fold-change (y-axis) over time in days (x-axis) for five clusters
identified by hierarchical clustering: 1 (red): monotonically
increasing, 2 (orange): increasing after initial decrease, 3 (dark
blue): monotonically decreasing, 4 (yellow): decreasing after initial
increase, 5 (light blue): non-monotonically decreasing. Each thin
transparent line represents one gene. Thick black-outline lines
represents the mean for each cluster. C Functional analysis of genes
with consistently increasing abundance (Cluster 1): (Top) Gene Ontology
(GO) enrichment analysis: Significantly enriched biological processes
(y-axis) and their key contributing genes (x-axis). Color intensity
indicates −log10(p value) of GO term enrichment. (Bottom) Gene ranking
based on MAGeCK analysis: Genes (x-axis) ranked by −log10(p value),
with color intensity indicating a more positive enrichment score. These
pathways may enhance proliferation when perturbed.
CRISPR long-term proliferation screen highlights TP53 as the top candidate
for knockout-mediated proliferation-enhancement
To investigate regulators that enable prolonged cell division and
bypass senescence, we performed a second CRISPR screen over 200 day
period, compared to the initial 30-day screen. bMSCs were transduced
with Cas9 and the sgRNA library at a higher multiplicity of infection
(MOI). Two control groups were included: untransduced cells and cells
transduced with Cas9 only. The total culture duration before
proliferation arrest increased from about 100 days in wild-type cells
to approximately 200 days in transduced cells, effectively extending
the proliferation period after controls stopped dividing (Fig. [99]4A).
Transduced cells also showed faster proliferation, with an average 12%
increase in doubling rate and up to a 50% increase by day 50
post-transduction compared to non-transduced controls (Fig. [100]4B).
To determine whether the decline in proliferation was associated with
senescence, we performed SA-β-gal staining at day 80, comparing
Cas9-transduced cells with the sgRNA library (+library) to those
without the library (−library) (Fig. [101]4C). The −library group
showed a significantly higher rate of senescence, with ~60% senescent
cells, compared to only ~10% in the +library group. This suggests that
the extended proliferative capacity at later time points in the
+library group is largely due to a reduced rate of senescence.
Fig. 4. Long-term proliferation and sgRNA dynamics in CRISPR-edited AD-bMSCs.
[102]Fig. 4
[103]Open in a new tab
AD-bMSCs were transduced with Cas9 and the sgRNA library, then cultured
for 200 days with DNA sampling every 5 days and RNA sampling at days 20
and 125. A, B Proliferation kinetics of sgRNA-library-transduced cells
(red, n = 3) compared to un-transduced (gray) and Cas9-only transduced
cells (black). Proliferation was measured by direct cell counting at
each passage. A Doubling time in hours over the course of the
experiment. B Cumulative number of doublings since the start of the
screen. C Boxplot showing the percentage of senescence-positive Cas9
+library cells compared to Cas9 −library cells at day 80. Non-library
cells had increased senescent rate compared to library cells. Each
point represents biological replicate (n = 3). Shaded areas indicate
the 95% confidence interval. D Relative abundance of sgRNAs over 100
days across three biological repeats: Top-left: Mean abundance of the
top 4 sgRNAs (red, blue, and green) and all other sgRNAs (gray).
Remaining panels: Individual abundance trajectories for each biological
repeat.
Sequencing DNA from multiple time points revealed significant shifts in
sgRNA abundance beginning at day 25, with TP53-targeting guides showing
a pronounced and consistent enrichment (Fig. [104]4D). In replicate C,
TP53 sgRNAs became nearly fixed in the population (>99%), while in
replicates A and B, they accounted for approximately 50%. The remaining
sgRNAs were primarily those targeting OSR2 in replicate B and a
non-cutting control (nc2480) in replicate A. Notably, OSR2 was
previously depleted in the short-term screen, and nc2480 is designed
not to impact cellular function. The mirrored enrichment of these
guides alongside TP53 is likely the result of co-inheritance due to
high multiplicity of infection (MOI) during lentiviral transduction,
rather than independent selective advantages. Thus, the apparent
enrichment of OSR2 and nc2480 likely reflects technical artifacts
rather than true biological effects. Despite these complexities, TP53
consistently emerged as the dominant and most robust candidate for
enhancing proliferation following knockout.
RNA sequencing of the CRISPR long-term proliferation screen identifies key
genes and pathways
To further understand the regulatory mechanisms governing AD-bMSCs
proliferation, we extracted and sequenced RNA from early wildtype (WT)
passage cells (day 25), late passage cells (day 125), and sgRNA
library-transduced cells (KO Library, day 125) collected on the
mentioned days of the second CRISPR screen (Fig. [105]4B). Principal
component analysis (PCA) revealed that late-passage KO library cells
(KL) retained similarity to early-passage WT cells (EWT), while
late-passage WT (LWT) cells diverged (Fig. [106]5A, see also Fig.
S[107]4), with this difference accounting for most of the variance
(63%). We examined the expression levels of known MSC markers and
candidate genes identified in the previous CRISPR screen
(Fig. [108]5B). Both LWT and KL cells expressed positive MSC markers at
levels similar to or higher than EWT, except for a lower expression of
CD29 in KL cells. Negative MSC markers were slightly elevated in KL
cells but remained low overall, indicating that KL cells retained some
MSC expression profile even at late passages. Differential expression
analysis using DESeq2 between all three groups identified
differentially expressed genes (DEGs, Fig. [109]5C). Specific
comparison of KL and LWT cells reveals genes that may contribute to KL
cells’ enhanced proliferation and younger expression profile
(Fig. [110]5D, see also Fig. S[111]5). Top DEGs include upregulated
genes like KCNK12, NREP, IDI1, and LTBP1, and downregulated genes such
as AQP1, SMOC1, SULF2, THBS1, and PLTP. Hierarchical clustering of gene
expression profiles revealed six distinct clusters (Fig. [112]5E).
Over-representation analysis of GO:BP terms for each cluster identified
cluster-specific pathways and terms (Fig. [113]5F). Key findings
included sterol and lipid metabolism for cluster 1 (up-regulated in
both KL and LWT), differentiation and cell adhesion terms for cluster 2
and 3 (downregulated in KL cells), DNA and cell-cycle terms for cluster
4 (upregulated in KL), and protein synthesis-related terms for clusters
5, 6 (downregulated in LWT cells). Our findings suggest that KL cells’
increased proliferation and extended culturing time may be attributed
to upregulated cell-cycle and DNA replication genes (cluster 4) and
maintained protein-synthesis gene expression levels similar to EWT
(Cluster 5). Despite these changes, KL cells downregulate
differentiation related genes, and specifically muscle-differentiation,
indicating that TP53 knockout might inhibit muscle-directed or
differentiation, or differentiation of MSCs in general.
Fig. 5. Transcriptomic analysis of wildtype and CRISPR-modified AD-bMSCs at
different passages.
[114]Fig. 5
[115]Open in a new tab
A Principal component analysis (PCA) of RNA-seq samples, showing KO
Library (day 125) cells clustering closer to WT (day 25) than to WT
(day 125). B Boxplot of normalized RNA-seq reads (y-axis) by cell type
(x-axis) for 8 selected genes, categorized into three groups: positive
MSC markers (red), negative MSC markers (blue), and candidate genes
from Fig. [116]4 (TP53 and OSR2, green). C Venn diagram of all
differentially expressed genes, grouped by group comparison. D Volcano
plot showing −log10(p value) vs. log2 fold change for genes
differentially expressed in KO Library (day 125) compared to Late WT
(day 125). Red points indicate enriched genes, blue points indicate
depleted genes. E Heatmap of differentially expressed genes (DEGs,
rows) across the three replicated of the three groups (columns). Values
represent variance stabilizing transformed (VST) and Z-score scaled
RNA-seq counts. DEGs are clustered into six groups (1–6) using
hierarchical clustering. F Top 4 enriched Gene Ontology (GO) Biological
Process category terms for each cluster. Y-axis: GO terms; X-axis: Fold
enrichment. Rows represent different clusters. Point size corresponds
to −log10(p value) of the enrichment.
Validation of TP53 and PTEN knockouts
To validate the results of the pooled CRISPR screen, TP53 and PTEN, the
top candidate genes, were individually knocked out in bMSCs using Cas9
and gene-specific sgRNAs, alongside a non-targeting sgRNA control.
Following transfection and 1 week of selection, two single-cell clones
from each knockout pool were isolated by serial dilution. Cell
viability, assessed by Alamar Blue assay (Fig. [117]6A), was
significantly higher in both TP53 and PTEN knockout (KO) populations
compared to wild-type (WT) cells (p < 0.01). Flow cytometry confirmed
that KO cells retained MSC surface marker expression (CD29, CD44) at
levels comparable to WT (Fig. [118]6B), while the non-targeting control
showed reduced expression, possibly due to off-target or sgRNA-related
effects. Genome editing efficiency, measured by sequencing and indel
analysis, reached 95% in TP53 KO and 43% in PTEN KO cells, with most
edits resulting in frameshifts (Fig. [119]6C). qRT-PCR further
confirmed strong knockdown of target gene expression in two independent
single-cell clones per gene (Fig. [120]6D). Adipogenic differentiation
was assessed via BODIPY (lipid) and Hoechst (nuclei) staining after 21
days in induction medium (Fig. [121]6E). The percentage of
BODIPY-positive cells was reduced in both knockouts dropping from 67.8%
in WT to 37.7% in TP53 KO and 25.5% in PTEN KO (Fig. [122]6F). However,
Hoechst-based total cell counts increased in three of the four KO
lines, with the fourth similar to WT (Fig. S[123]6), suggesting that
the reduced proportion of adipogenic cells may partly reflect ongoing
proliferation, which increases total cell number, rather than a
complete loss of differentiation capacity. Senescence, assessed by
SA-β-gal staining (Fig. S[124]6), was comparable between WT and
knockouts. Notably, TP53 KO cells exhibited a higher absolute number of
cells, consistent with their overall increase in cell number. In
summary, targeted knockout of TP53 or PTEN enhances bMSC proliferation
without increasing senescence, but reduces adipogenic differentiation
efficiency. While this tradeoff may not preclude their application in
cultured meat, further optimization is needed to balance expansion and
differentiation potential. These results validate the CRISPR screen and
underscore the utility of gene editing to improve MSC-based cultured
meat production.
Fig. 6. Validation of TP53 and PTEN knockouts in AD-bMSCs.
[125]Fig. 6
[126]Open in a new tab
AD-bMSCs were transfected with Cas9 and an sgRNA targeting either TP53
(red), PTEN (green), or a non-exonic targeting control sgRNA (gray)
(n = 3 per group). A Cell viability was measured using the Alamar Blue
assay at days 3, 5, and 7 after seeding. Both TP53 and PTEN knockout
cells exhibited significantly increased viability compared to wild-type
(WT) controls. Shaded areas indicate the 95% confidence interval. B
Flow cytometry analysis of surface markers CD29 (APC) and CD44 (FITC)
in bMSCs. TP53 and PTEN knockout cells maintained surface marker
expression levels similar to WT cells, while the control group showed a
minor reduction. C Editing efficiency and indel distribution in
knockout populations, as determined by next-generation sequencing. Bars
represent the percentage of sequences with specific insertions (+) or
deletions (−). TP53 knockout cells showed 95% editing efficiency, while
PTEN knockout cells showed 43%. D qRT-PCR analysis of TP53 and PTEN
mRNA expression in knockout and WT cells, normalized to the RBPs
housekeeping gene. Two independent single-cell clones (SC) for each
knockout showed significant reductions in target gene expression. E
Representative fluorescence microscopy images showing adipogenic
differentiation after 21 days in induction medium. Lipid droplets were
stained with BODIPY (green), and nuclei with Hoechst (blue). F
Quantification of adipogenic differentiation efficiency, expressed as
the percentage of BODIPY-positive cells. Both TP53 and PTEN knockout
clones showed a significant reduction in adipogenic potential compared
to WT. Statistical significance is indicated (*p < 0.05, **p < 0.01,
***p < 0.001).
Discussion
Slow proliferation and early senescence limit the scalability of bovine
mesenchymal stem cells (bMSCs) for cultured meat production. To tackle
those challenges, we conducted a CRISPR-based functional genomics
screen in bMSCs, targeting proliferation with a library of 603 gene
knockouts. We discovered that inactivating TP53 and PTEN genes most
significantly enhanced bMSC proliferation and delayed cellular
senescence. These results directly address slow cell expansion as a
major obstacle in cultured meat production by providing a genetic
approach to shorten timelines and lower costs for industrial-scale
manufacturing^[127]16. The prominence of these well-established tumor
suppressors as top hits provides a foundation for cell line engineering
but necessitates careful consideration of the mechanisms and safety
implications of their manipulation.
Specifically, TP53 knockout significantly increased proliferation,
maintained early-passage gene expression, and extended population
doublings by delaying senescence onset. This aligns with TP53’s
well-established role as a key regulator of senescence, apoptosis, and
DNA repair^[128]33, as well as its frequent identification in CRISPR
knockout screens across diverse cell lines^[129]36. Proliferation is
thus enhanced via two complementary mechanisms: accelerating cell
division to produce more cells in less time, and postponing senescence
to prolong the proliferative lifespan, resulting in a greater overall
cell yield. Similarly, PTEN knockout emerged as a strong candidate due
to its involvement in cell cycle regulation and apoptosis^[130]37,
although its effects on senescence appear context-dependent^[131]34.
Nonetheless, these proliferative gains often come with trade-offs in
differentiation potential.
Balancing proliferation with differentiation potential is therefore
essential for optimizing MSCs for cultivated meat applications^[132]38.
Although prior studies have linked TP53 knockout to enhanced adipogenic
differentiation^[133]39–[134]41, we found that both TP53 and PTEN
knockouts reduced the proportion of adipogenic cells. However, the
total number of adipogenic cells was less affected, likely because
undifferentiated cells with a proliferative advantage outcompeted
differentiating ones. Similarly, TP53 knockout populations showed
downregulation of muscle-related genes, suggesting diminished myogenic
potential^[135]21 and posing limitations for cultured muscle
production. While this study focuses on adipogenic-competent MSCs for
fat tissue applications, cell lines intended for myogenic
differentiation will require independent validation to ensure that
proliferation-enhancing modifications do not impair muscle development
and function.
Beyond TP53 and PTEN, our screen uncovered additional gene knockouts
that may promote proliferation through distinct metabolic and signaling
mechanisms. Knockout of mitochondrial pyruvate carrier^[136]42 (MPC1
and MPC2) subunits increased cell abundance, suggesting proliferative
advantages through metabolic shifting toward aerobic
glycolysis^[137]43,[138]44 a shift associated with mitochondrial
involvement^[139]45,[140]46. Similarly, knockouts of SMAD3, SMAD4, and
SMAD5 increased bMSC abundance, highlighting their involvement in TGF-β
and BMP signaling pathways that govern chondrogenesis^[141]47. Notably,
disrupting chondrogenesis-associated genes (e.g.,: SMADs, RUNX1^[142]48
and SERPINH1^[143]49) may accelerate proliferation while reducing
undesired cartilage differentiation, providing dual benefits for
cultivated meat production.
Building on these findings, this CRISPR screening framework can be
readily adapted to other phenotypes^[144]50 and cell sources^[145]17
critical for cultivated meat production, including satellite cells,
embryonic stem cells, and induced pluripotent stem cells, which can be
combined to produce cultured meat products. Future screens could
interrogate adipogenic and myogenic differentiation, reduced serum
dependence, and growth in suspension culture, identifying novel
regulators of lineage commitment and metabolic adaptation.
The use of CRISPR-edited cells in food production raises important
safety and regulatory challenges. Extended culturing of cells with
tumor-suppressor knockouts may induce genomic instability,
necessitating careful monitoring^[146]51. For large-scale production,
transient gene modulation using chemical inhibitors or inducible
systems may be safer than permanent genetic edits. However, only
compounds with established safety profiles (e.g., GRAS status in the
U.S. or EFSA approval in Europe) are likely suitable for food
applications. While CRISPR does not necessarily introduce foreign DNA,
regulatory frameworks in many countries consider it as genetic
modification, requiring strict oversight and potentially facing
consumer skepticism^[147]52. Future implementation will require careful
selection of approved compounds, validation of their removal from
products, and transparent safety testing to ensure consumer trust and
regulatory compliance across global markets.
Future research should prioritize evaluating both the safety and
functionality of these modified bMSCs in food systems, while also
exploring combinatorial approaches to further improve efficiency. The
versatility of CRISPR technology offers opportunities to systematically
optimize cell lines for scalable biomanufacturing, with
interdisciplinary collaboration being crucial for advancing sustainable
alternatives to conventional meat production.
Online methods
Bovine MSC isolation and characterization
This section outlines the isolation of adipose-derived bovine MSC
(AD-bMSCs) from calf tissue, their culture methods, and
characterization procedures. We describe the process for cell
isolation, the expansion and maintenance of AD-bMSCs in culture, and
the analysis of their growth kinetics. Additionally, we detail the flow
cytometry methods used to confirm MSC-specific surface markers, and the
differentiation assays employed to verify multipotency. These
procedures collectively establish the identity and quality of the
AD-bMSCs used in subsequent CRISPR screening experiments.
AD-bMSCs isolation
AD-bMSCs were isolated from 3-month-old female calves immediately
post-slaughter, following a modified protocol from Sampaio et
al.^[148]27. Briefly, adipose tissue was aseptically collected and
washed with phosphate-buffered saline (PBS; Biological Industries,
Israel) containing Pen-Strep Solution (Biological Industries). The
tissue was surface sterilized with 75% ethanol for 30 s. Blood vessels
and connective tissues were meticulously removed, and the adipose
tissue was minced into small pieces before being digested with
Collagenase type I solution (2 mg/mL in DMEM/LG; Worthington, USA) at a
1:10 ratio for 3 h at 37 °C. The enzymatic reaction was terminated by
adding an equal volume of Dulbecco’s Modified Eagle’s Medium Low
Glucose (DMEM/LG; Biological Industries) supplemented with 15% fetal
bovine serum (FBS; Biological Industries). The digested product was
filtered through a 200-µm mesh sieve and centrifuged at 300 × g for
10 min. The resulting cell pellet was resuspended in culture medium and
seeded in 10 cm² dishes. Cells were incubated at 37 °C in a 5% CO₂
atmosphere. After 24 h, non-adherent cells were removed by washing with
PBS. When the cells reached 80% confluence, they were passaged using
trypsin (Biological Industries, Israel).
AD-bMSCs routine culturing
AD-bMSCs were expanded in low-glucose DMEM supplemented with 15% fetal
bovine serum (FBS), 2 mM L-glutamine, and 1% Pen-Strep X100 (final
concentrations: 100 units/mL Penicillin G Sodium Salt, 0.1 mg/mL
Streptomycin Sulfate) (Sartorius). Cells were maintained in a
humidified incubator at 37 °C with 5% CO[2] and passaged every 3–4 days
at 60–80% confluency using 0.25% Trypsin Solution A (Sartorius). For
cryopreservation, cells were counted, centrifuged at 400 × g, and
resuspended in full growth medium containing 10% DMSO v/v
(Sigma-Aldrich, Israel). Cryovials were initially stored at −80 °C for
short-term preservation, then transferred to liquid nitrogen for
long-term storage.
Cell culture and growth kinetics
Growth kinetics were analyzed for AD-MSCs from thirteen different
replicates, spanning passages 3 through 25. Cells were seeded in
multi-well plates or 10 cm cell dishes at a density of approximately
7000 cells/cm^2 and maintained in MSC culture medium at 37 °C in a
humidified atmosphere with 5% CO[2]. The medium was refreshed every
48 h. Cells were passaged and counted every 96 h (coinciding with every
other medium replacement). Briefly, cells were trypsinized, stained
with trypan blue, and viable cells were counted using a Countess II
hemocytometer (Thermo Fisher Scientific, USA). Population doubling time
(Td) during the logarithmic growth phase was calculated using the
Patterson formula:
[MATH: Td=t×lg2/lgnt−lgNo :MATH]
Where: t = incubation time, No = initial cell number, Nt = cell number
at time t.
Senescence-associated β-gal (SA-β-gal) staining
To assess the percentage of senescent cells in AD-bMSCs, SA-β-gal
staining was performed using the Senescence β-Galactosidase Staining
Kit (Cell Signaling Technology, #9860), following the manufacturer’s
instructions. Briefly, cells were seeded at a density of 1 × 10⁴
cells/cm². After 24 h, 20 µM etoposide (Abcam, AB-ab120227-25-B) was
added to the relevant treatment groups. Cells were subsequently washed
once with PBS and fixed with the provided Fixative Solution for 15 min
in a chemical hood. Following fixation, cells were washed twice with
PBS and incubated with the Staining Solution. Plates were sealed with
parafilm and incubated in a dry 37 °C incubator (no CO₂) for at least
24 h. After staining, cells were imaged, and senescent cells were
manually annotated and counted.
Differentiation assays
Adipogenic differentiation
AD-bMSCs were subjected to adipogenic differentiation for 14 days using
an alternating induction and maintenance medium protocol. The induction
medium consisted of Low-glucose DMEM supplemented with 10% FBS, 1%
penicillin/streptomycin, 1% L-glutamine, 1 μM dexamethasone, 0.5 mM
3-isobutyl-1-methylxanthine, 0.2 mM indomethacin, and 10 μg/mL insulin
(all from Sigma-Aldrich). Cells were cultured in this medium for 3
days, followed by 1 day in maintenance medium (DMEM-LG, 10% FBS, 1%
penicillin/streptomycin, and 10 μg/mL insulin). This cycle was repeated
for the duration of the experiment. Control cells were maintained in
standard growth medium. After 14 days, adipogenic differentiation was
quantified via lipid content as assessed by either Oil Red O (ORO)
staining or BODIPY staining (Invitrogen). For ORO staining, cells were
washed twice with PBS and fixed with 4% formaldehyde for 10 min at room
temperature. Fixed cells were then incubated for 1 h with freshly
prepared ORO working solution (6 parts 0.5% ORO stock solution in
isopropanol mixed with 4 parts distilled water). Cells were then washed
thoroughly with distilled water before imaging. For BODIPY imaging,
cells were washed twice with PBS, and then imaged.
Chondrogenic differentiation
AD-bMSCs at passages 9 were induced to undergo chondrogenic
differentiation for 21 days. The chondrogenic medium (CM) consisted of
DMEM/F12 supplemented with 10% FBS, 1% insulin-transferrin-selenium
(ITS), 50 μg/mL proline, 0.1 μM dexamethasone, 0.9 mM sodium pyruvate,
50 μg/mL L-ascorbic acid, and 10 ng/mL transforming growth factor β3.
The culture medium was replaced every 3 days. Chondrogenic
differentiation was evaluated by Alcian Blue staining to detect
glycosaminoglycan (GAG) production. After 21 days, cells were fixed in
cold methanol for 5 min, then incubated in 1% Alcian Blue solution (pH
2.5) for 30 min at room temperature. Cells were then washed three times
with PBS before imaging.
Flow cytometry analysis
Flow cytometry was performed on AD-bMSCs at different passages to
investigate the expression of MSC-specific markers CD29 and CD44.
AD-bMSCs were harvested using 0.25% trypsin-EDTA. The cell suspension
was centrifuged at 300 × g for 5 min and washed three times with DPBS
containing 1% bovine serum albumin (Sigma-Aldrich). The AD-bMSCs were
filtered through a 70-μm nylon cell strainer, and cell concentration
was determined using a hemocytometer. Approximately 5 × 10^5 cells per
sample were labeled with FITC-conjugated anti-bovine CD44 antibody
(1:250 dilution, Catalog #AB_528148, DSHB) and incubated at 4 °C in the
dark for 30 min. In parallel, an equal number of cells were directly
labeled with APC-conjugated anti-bovine CD29 antibody (1:250 dilution,
Catalog #303008, BioLegend). All samples were washed twice with PBS and
centrifuged at 300 × g at 4 °C for 5 min. The washed cells were
resuspended in 200 μL of PBS and filtered through a 100-μm nylon cell
strainer. Labeled samples were analyzed using a BD Accuri™ C6 Plus Flow
Cytometer. For each sample, at least 10,000 events were recorded. Data
analysis and plot generation were performed using R with packages
flowCore and ggcyto (see “Bioinformatics” section for details).
Positive expression was defined as fluorescence intensity greater than
99% of the corresponding unstained control.
Cas9 mRNA expression validation in transduced cells
RNA was extracted and amplified from Cas9-transduced cells to verify
mRNA expression. Total RNA was isolated from 5 × 10⁶ cells using TRIzol
reagent (Invitrogen), followed by purification with the Direct-zol RNA
Miniprep Plus Kit (Zymo Research). cDNA was amplified using Taq Mix Red
(PCRBIO) according to the manufacturer’s instructions. PCR products
were run on a 2% agarose gel and visualized using a UV
transilluminator. Primer sequences are found in Table [149]S1.
Reporting summary
Further information on research design is available in the [150]Nature
Portfolio Reporting Summary linked to this article.
CRISPR library construction and screening
This section outlines the design and construction of a custom CRISPR
library targeting key genes involved in cell cycle regulation, cloned
into a lentiviral vector system that co-expressed Cas9. Following
lentiviral production and transduction, the bMSCs population was
cultured under proliferative conditions. Genomic DNA was extracted at
various timepoints for next-generation sequencing (NGS) analysis,
enabling the identification of genes whose knockout enhanced bMSCs
proliferation while preserving differentiation capacity.
CRISPR library design
We designed a focused CRISPR knockout library targeting 603 Bos taurus
genes related to proliferation, including transcription factors and
genes selected from the Dependency Map (DepMap) portal based on their
Gene Effect Scores. The library comprised 2564 sgRNAs targeting coding
regions and 436 control sequences. sgRNA sequences were retrieved and
filtered from a previously published whole-genome CRISPR library, with
NGS, barcoding adapters, and restriction enzyme recognition site added
to the sequences. Detailed information on gene selection criteria and
library composition is provided in the “Results” section.
Library oligonucleotide synthesis, amplification and cloning
The custom CRISPR library was synthesized as an oligonucleotide pool by
Twist Biosciences. Upon receipt, the lyophilized pool was reconstituted
in TE buffer to a concentration of 1 ng/μL. Single-stranded oligos were
PCR-amplified using universal library adapter primers in triplicate.
The amplified products were combined, purified using the Promega
Wizard® SV Gel and PCR Clean-Up System, and verified on a 2.5% agarose
gel. Concentration was determined using both NanoDrop and Qubit
instruments. The amplified library was inserted into LentiGuide-Puro
(Addgene #52963) using a GoldenGate reaction with BsmbI/Esp3I enzymes,
following a modified protocol based on Read et al. Oligonucleotide
sequences are found in Table [151]S2.
Bacterial transformation and library amplification
Ligation products were purified by ethanol precipitation and
resuspended in 5 μL TE buffer. One microliter of purified ligation
product was electroporated into 50 μL of either NEB Stable or Lucigen
Endura electrocompetent cells using a Gene Pulser Xcell Electroporation
System (Bio-Rad) at 1800 V, 600 Ω, and 10 μF. post-electroporation,
bacteria were recovered in 1 mL SOC medium for 1 h at 37 °C with
250–300 RPM shaking. Each 1 mL of recovered bacteria was split 10%:90%;
the 10% fraction was used for serial dilution plating and colony
counting, while the 90% fraction was amplified for plasmid library
extraction. Colony forming units (CFUs) were determined by plating
dilutions of 1:2 × 10^5, 1:2 × 10^6, and 1:2 × 10^7 on
ampicillin-containing agar plates. Library coverage was calculated,
with a minimum threshold of 1000 CFUs per sgRNA. For plasmid library
amplification, the 90% fraction of recovered bacteria was diluted into
250 mL LB + ampicillin and grown at 25 °C and 250–300 RPM for at least
16 h until OD ≥ 1. Approximately 10 aliquots of 1 mL bacteria were
cryopreserved with 25% glycerol. The remaining culture was used for
plasmid DNA extraction using “NucleoBond Xtra Midi” (MACHEREY-NAGEL),
with final elution in 500 μL. Plasmid concentration and purity were
determined by NanoDrop.
Lentiviral production
The following lentivirus-production related plasmids were amplified in
bacteria and extracted using “NucleoBond Xtra Midi” as per
manufacturers instruction: Second-generation lentiviral envelope and
packaging plasmids (VSV-G envelope pMD2.G, Addgene #12259; psPAX2,
Addgene #12260), a control lentiGFP plasmid, and two CRISPR-Cas9
plasmids (Lenti-iCas9-neo, Addgene #85400; lentiCas9-Blast, Addgene
#52962) were amplified in bacteria and extracted using standard plasmid
midi-prep protocols. For transfection, a mix of VSV-G/psPAX2/Transfer
vector was prepared at a ratio of 3:1:4 and introduced into HEK293T
cells using TurboFect (Rhenium) reagent as per manufacturers
instruction. The medium was replaced with complete HEK293T medium 24 h
post-transfection. Lentiviral particles were harvested 48–72 h
post-transfection by collecting the cell culture medium and
centrifuging at 400 × g at 4 °C for 5 min to remove cellular debris.
The supernatant was supplemented with 10 mM HEPES buffer to stabilize
pH, filtered through a 0.45 μm PES filter, aliquoted, and snap-frozen
in liquid nitrogen before storage at −80 °C.
Lentiviral transduction optimization via flow cytometry analysis
MSCs were transduced with GFP vector-containing lentivirus to assess
transduction efficiency. Cells were seeded in 6-well plates at a
density of 1.5 × 10^5 cells per well 24 h prior to transduction, or on
the day of transduction for reverse transduction. Lentiviral particles
were added to the cells at various multiplicities of infection (MOI),
in the presence of different transduction enhancers (8 or 4 μg/mL
polybrene, 10 or 5 μg/mL protamine sulfate, or 1X ViralEntry). Both
forward and reverse transduction methods were tested. For reverse
transduction, virus and enhancers were mixed with cells upon seeding.
Cells were incubated with virus for 24 h, after which the media was
replaced with fresh growth media. GFP expression was analyzed 72 h
post-transduction by splitting and washing the cells, then measuring
GFP intensity using a BD Accuri C6 flow cytometer. A minimum of 10,000
events were collected for each sample. Data were analyzed using R
flowCore package to determine the percentage of GFP-positive cells and
mean fluorescence intensity.
Lentiviral transduction and titration
Transduction was performed using a reverse transduction method. Cells
were passaged, counted, and resuspended in fresh low-glucose DMEM
supplemented with varying concentration of transduction enhancers
(Polybrene, Protamine sulfate, or VirlaEntry). Lentivirus was rapidly
thawed at 37 °C, added to the cell suspension, gently mixed, and plated
on 10 cm cell culture dishes. After 48 h, the medium was replaced with
antibiotic-containing medium for selection. Selection regimes varied by
antibiotic: puromycin (2.5 μg/mL, 3–4 days), blasticidin (5 μg/mL, 5–7
days), or G418 (1 mg/mL, 7–10 days). Selection continued until all
negative control untransduced cells died.
To determine lentiviral titer and transduction efficiency, a serial
dilution assay was performed. Cells were seeded in a 6-well plate with
varying amounts of lentivirus (500 μL, 250 μL, 125 μL, 62 μL, and
32 μL) plus two control wells without lentivirus. Selection antibiotic
was added to all wells except one control. This setup allowed for the
determination of 100% cell survival (control without selection) and 0%
transduction (control with selection). Transduction efficiency was
calculated by comparing the number of surviving cells in each dilution
to the 100% survival control, after subtracting the number of cells in
the 0% transduction control:
[MATH: TEN
(%)=[
(NCN-NC0%<
/mi>)/NC100%<
/msub>]/VolN
:MATH]
Where: TE = transduction efficiency in percentage, Vol[N] = Lentivirus
volume in well N, NC = number of counted cells at well N, well 0% (no
virus with antibiotics selection), well 100% (no virus with no
antibiotic selection). This method enabled the calculation of
transduction efficiency as a percentage and the determination of
functional titer in transduction units per mL.
Targeted CRISPR screening
Two CRISPR screens were conducted to identify genetic factors
influencing bMSCs proliferation.
30-day screen
Cas9-expressing bMSCs were transduced with a lentiviral sgRNA library
at a low multiplicity of infection (MOI ~ 0.3) to ensure approximately
600 cells per sgRNA. The screen was conducted over a 30-day period,
with cells passaged every 3–4 days. DNA samples were collected at
specified intervals for sgRNA quantification via next-generation
sequencing (NGS). sgRNA quantification, ranking, and statistical
testing were performed using MAGeCK software.
200-day screen
A second, extended CRISPR screen was performed over a 200-day
timeframe. bMSCs were initially cultured for 75 days (approximately 100
doublings) before being transduced with Cas9 and the sgRNA library at a
higher MOI to assess combinatorial gene knockout effects. Two control
groups were included: untransduced cells and cells transduced only with
Cas9. Cells were passaged and sampled as in the 30-day screen, with DNA
extraction and NGS performed at specified timepoints. Data analysis was
conducted using MAGeCK as in the 30-day screen. For both screens,
lentiviral particles were produced as described in the “Lentivirus
creation and transduction” section. Transduction efficiency was
determined using the titration method outlined previously. Cell culture
conditions and passaging protocols were consistent with those described
in the “AD-bMSCs routine culturing” section.
DNA extraction and amplification for NGS
Genomic DNA (gDNA) was extracted from 5 × 10^5 cells using QuickExtract
DNA Extraction Solution (Lucigen) following the manufacturer’s
protocol. The extracted gDNA underwent two rounds of PCR amplification.
The first PCR amplified the sgRNA-containing region, and the product
was purified using AMPure XP size selection beads (Beckman Coulter).
The second PCR added NGS adapters to the amplicons, followed by another
round of purification using two-sided size selection with AMPure XP
beads. The final amplicon concentration was determined using a Qubit
fluorometer, and the product size was verified on a 2% agarose gel.
sgRNA amplicons were then sequenced using the xGen™ Broad-Range RNA
Library Preparation Kit (IDT) on an Illumina platform at IDT/Syntezza.
RNA sequencing
Total RNA was extracted from 5 × 10^6 cells using TRIzol reagent
(Invitrogen) followed by Direct-zol RNA Miniprep Plus Kits (Zymo
Research). RNA quality was assessed using an Agilent 2200 TapeStation
Bioanalyzer, and all samples had RNA Integrity Number (RIN) > 8.5.
Library from total RNA was prepared using MARS-seq protocol^[152]53,
sequenced on an Illumina NovaSeq 6000 sequencer with single-end reads
of 75 bp, and aligned to bovine reference genome (ARS-UCD1.2) at The
Nancy and Stephen Grand Israel National Center for Personalized
Medicine (G-INCPM).
CRISPR-Cas9 knockout of PTEN and TP53 sgRNA design and synthesis
sgRNAs targeting PTEN and TP53 were designed using the sgRNA Scorer 2.0
platform and synthesized from complementary oligonucleotides by
annealing. sgRNAs were cloned into the lentiCRISPR v2 backbone plasmid
using Golden Gate Assembly (Esp3I digestion). Plasmids were transformed
into E. coli, selected on ampicillin plates, and verified by colony
PCR. Positive clones were isolated by miniprep (Tiangen) and confirmed
by Sanger sequencing. Single guide RNA sequences, as well as the
primers used to construct them are found in Tables [153]S3 and [154]S4,
respectively.
Cell culture and transfection
AD-bMSCs were seeded in 6-well plates 24 h prior to transfection. For
each well, 4 μg plasmid was mixed with 400 μl Opti-MEM and 6 μl
TurboFect reagent. After 20 min, the mixture was added to cells for 4 h
before replacing with DMEM. Transfected cells were selected with
2.5 μg/ml puromycin for 96 h, starting 72 h post-transfection.
TP53 and PTEN KO DNA editing validation of knockout cell lines
To validate CRISPR-Cas9 editing of the target genes, genomic DNA was
extracted from selected cells using QuickExtract DNA Extraction
Solution (Lucigen). PCR was performed using primers flanking the
targeted regions. Products were purified with SPRI beads and analyzed
by agarose gel electrophoresis. The resulting amplicons were sequenced
using Sanger sequencing (Hylabs, Israel). Sequencing traces were
analyzed using the TIDE (Tracking of Indels by DEcomposition)
bioinformatics tool to quantify editing efficiency and characterize the
induced mutations. Primers for PTEN and TP53 knockout validations are
found in Table [155]S5.
TP53 and PTEN KO qRT-PCR knockout validation
Cells were seeded in triplicate and cultured until they reached 80%
confluency. At this point, cells were trypsinized and immediately fixed
with TRI-Reagent (MRC, USA). Total RNA was then extracted using the
Direct-zol RNA MiniPrep Plus kit (Zymo Research, USA), following the
manufacturer’s instructions. The concentration and purity of the
extracted RNA were assessed using NanoDrop One spectrophotometer
(Thermo Scientific). For cDNA synthesis, 400 ng of total RNA was
reverse transcribed using the UltraScript cDNA Synthesis Kit (PCR
BioSystem, USA) according to the manufacturer’s protocol. Quantitative
PCR (qPCR) was performed using the qPCRBIO SyGreen Blue Mix Hi-ROX (PCR
BioSystem, USA) on a StepOnePlus Real-Time PCR system (Applied
Biosystems). Specific primers for genes of interest and housekeeping
genes are listed in Table [156]S5. Gene expression levels were
normalized to the housekeeping gene RBPs. Relative quantification of
gene expression was calculated using the 2 − ∆∆Ct method. All qPCR
reactions were performed in technical triplicates for each biological
replicate. Primer design and sequences can be found in Table [157]S6.
Alamar blue cell proliferation assay
Cell proliferation was assessed using the Alamar Blue assay. Cells were
seeded in 96-well plates (1000 cells/well, six replicates) and measured
on days 3, 5, and 7. For each measurement, the medium was replaced with
90 μl fresh medium and 10 μl Alamar Blue reagent. After 3 h incubation
at 37 °C, absorbance was measured at 540 nm.
Bioinformatic analysis
All data analysis was performed in R (v4.2.2), using tidyverse packages
for general data tidying or other specialized packages for specific
analysis, as detailed below. All the graphs were generated with
ggplot2.
Flow cytometry
Data analysis and plot generation were performed the flowCore and
ggcyto packages. Cells were gated using a two-step process: first based
on forward and side scatter properties to exclude debris and doublets
(FSC-A > 3 × 10⁶ and SSC-A < 7.5 × 10⁶ and exclusion of maximum FSC-A
events), followed by selection of the normally distributed population
using the norm2Filter function from the flowCore package. Positive
expression was defined as fluorescence intensity greater than 99% of
the corresponding unstained control cells. CD44 and GFP expression were
analyzed using FITC fluorescence, and CD29 expression was analyzed
using APC fluorescence.
CRISPR screen
CRISPR screen data analysis was conducted using custom R scripts
alongside established bioinformatics tools. Raw sequencing reads were
first normalized with DESeq2 to account for differences in sequencing
depth across samples. sgRNA enrichment analysis was then performed
using MAGeCK (Model-based Analysis of Genome-wide CRISPR-Cas9
Knockout), which applies a negative binomial model to detect
significantly enriched or depleted sgRNAs between conditions. These
sgRNA-level results were aggregated into gene-level scores using the
robust rank aggregation (RRA) algorithm implemented in MAGeCK. To
ensure data quality, we evaluated the distribution of read counts,
replicate correlations, and the expected behavior of positive and
negative control genes. For RNA-seq data associated with the screen,
differential gene expression analysis was carried out using DESeq2.
Genes were ranked based on MAGeCK scores and false discovery rates
(FDR), and hits were prioritized using combined thresholds for fold
change and statistical significance.
Gene ontology over-representation analysis
Gene ontology (GO) over-representation analysis was conducted to
identify enriched biological processes associated with significant
changes in sgRNA abundance over time. Log2 fold changes in sgRNA
abundance across multiple time points (days 0, 3, 7, 10, 16, 24, and
30) were used to cluster genes via hierarchical clustering with
complete linkage based on Spearman correlation coefficients, resulting
in five clusters (Fig. S[158]3A, B). GO enrichment analysis was
performed using the g:Profiler2 R package (gprofiler2), applying the
gost function separately for each cluster. Genes were ranked by their
maximum absolute log2 fold change values. Enriched GO terms were
filtered to retain those unique to each cluster. Additional statistics
calculated included term ratio, total ratio, log2 fold enrichment, and
−log10(p value). To reduce redundancy among enriched terms, semantic
similarity was assessed using the rrvgo package with a threshold of
0.7. The most significant term was selected as the representative for
each group of similar terms.
RNA-seq
Differential expression analysis was conducted using DESeq2 in R
(v4.2.2). Genes with an adjusted p value < 0.05 and |log2 fold
change | > 1 were considered significantly differentially expressed.
Gene Ontology (GO) enrichment analysis was performed using the
clusterProfiler R package to identify significantly enriched biological
processes, molecular functions, and cellular components (adjusted p
value < 0.05), followed by results reduction via rrvgo R package.
Use of large language models and illustration tools
Large language models (LLMs) were employed to assist in editing and
proofreading the manuscript. All LLM-generated content was subsequently
reviewed and validated by the authors to ensure accuracy and prevent
errors or hallucinations. Figure [159]1A was created using a paid,
licensed BioRender account.
Statistics and reproducibility
All statistical analyses were performed in R (v4.2.2). Welch’s t tests
were used where appropriate. All post hoc comparisons employed Tukey’s
honestly significant difference (HSD) test unless otherwise noted.
Effect sizes were reported as Cohen’s d with 95% confidence intervals
(CI). Significance thresholds were defined as follows: p < 0.05 (*),
p < 0.01 (**), and p < 0.001 (***).
Senescence and marker expression (Fig. [160]1)
Senescence analysis (Fig. [161]1C) showed a significant effect of
treatment on the percentage of senescent cells, as determined by
one-way ANOVA (F[2,42] = 492.82, p < 0.001, η² = 0.96, 95% CI [0.94,
1.00]). Post hoc Tukey’s HSD indicated that late-passage cells had
higher senescence than early-passage cells (mean difference = 23.27,
p < 0.001). Etoposide-treated early-passage cells showed increased
senescence versus both early-passage (76.11, p < 0.001) and
late-passage cells (52.84, p < 0.001).
Flow cytometry surface marker expression (Fig. [162]1E, n = 3 per
group) showed increased CD44 in MSCs compared to control (mean
difference = 1.64, 95% CI [1.59, ∞], t(2.36) = 32.39, p < 0.001,
Cohen’s d = 26.44, 95% CI [7.17, ∞]) and increased CD29 (1.71, 95% CI
[1.62, ∞], t(2.14) = 22.79, p < 0.001, d = 18.61).
Lentiviral transduction efficiency (Fig. [163]1F) was significantly
higher in MSCs (mean difference = 0.92, 95% CI [0.62, ∞],
t(2.57) = 7.70, p = 0.004, d = 6.44) and HEK293T cells (1.69, 95% CI
[1.65, ∞], t(2.09) = 143.41, p < 0.001, d = 117.54) relative to
controls.
CRISPR screen and senescence validation (Figs. [164]2 and [165]4)
Modified hypergeometric (mHG) tests (Fig. [166]2C and Table [167]S7)
detected significant enrichment and depletion of sgRNAs across
timepoints. TGES guides count significantly increased as early as day
3, while BGES guides count decreased by day 7 with sustained enrichment
through day 30. N = 603, n_max = 603, B[(TGES)] = 73, B[(BGES)] = 46
for all days. Other parameters are found in Table [168]S7A, B.
Senescence was increased in the Cas9+Library group compared to
Cas9–Library (Fig. [169]4C, n = 3), with a mean difference of 51.98,
95% CI [44.43, 59.53], t(4) = 19.11, p < 0.001, d = 19.12.
qPCR and viability analyses (Fig. [170]6A–D)
Alamar Blue viability assays (Fig. [171]6A, n = 3 per group) were
analyzed using a linear mixed-effects model including fixed effects for
treatment and day, and a random intercept for replicate. Viability was
significantly increased in PTEN KO (+5893, SE = 2165, t(13) = 2.72,
p = 0.017), TP53 KO SC1 (+14,675, SE = 2163, t(21) = 6.79, p < 0.001),
and TP53 KO SC2 (+17,011, SE = 2163, t(21) = 7.87, p < 0.001). No
difference was observed between SC1 and SC2 (p = 0.54).
RT-qPCR (Fig. [172]6D, n = 3) revealed that TP53 expression was reduced
in SC1 (mean difference = 8.62, 95% CI [7.99, 9.25], p < 0.001) and SC2
(9.27, 95% CI [8.64, 9.90], p < 0.001) vs. wild type, with SC2 slightly
lower than SC1 (0.65, 95% CI [0.02, 1.28], p = 0.046). PTEN expression
was lower in SC1 (0.97, 95% CI [0.39, 1.54], p = 0.005) and SC2 (1.34,
95% CI [0.77, 1.91], p < 0.001) vs. wild type, with no significant
difference between SC2 and SC1 (0.38, 95% CI [−0.20, 0.95], p = 0.188).
Surface marker analysis via flow cytometry (Fig. [173]6B, n = 3) showed
CD29 expression was elevated in P53 KO (mean difference = 1.24, 95% CI
[0.76, 1.72], t(2) = 11.18, p = 0.008) and PTEN KO (1.64, 95% CI [1.39,
1.90], t(2) = 27.93, p = 0.001), but not in NC74 (0.87, 95% CI [−0.13,
1.88], t(2) = 3.75, p = 0.064). CD44 was increased in NC74 (1.02, 95%
CI [0.47, 1.57], t(2) = 7.98, p = 0.015), P53 KO (1.36, 95% CI [0.89,
1.84], t(2) = 12.34, p = 0.007), and PTEN KO (1.56, 95% CI [1.44,
1.66], t(2) = 58.96, p < 0.001) relative to control.
Adipogenic differentiation (Fig. [174]6F, n = 3 per group) was reduced
in TP53 KO SC1 (mean difference = 40.35%, SE = 5.38, t(10) = 7.50,
p < 0.001) and SC2 (19.96%, t(10) = 3.71, p = 0.026), as well as PTEN
KO SC1 (57.67%, t(10) = 10.73, p < 0.001) and SC2 (27.00%,
t(10) = 5.02, p = 0.0037), compared to wild type.
Supplementary information
[175]Supplementary Information^ (2.1MB, pdf)
[176]Supplementary Data^ (2.3MB, xlsx)
[177]42003_2025_8760_MOESM3_ESM.pdf^ (40.4KB, pdf)
Description of Additional Supplementary Files
[178]Reporting Summary^ (6.6MB, pdf)
Acknowledgements