Nutlin-3 – DMF inhibitor

Sequential cancer mutations in cultured human intestinal stem cells

Jarno Drost1,2, Richard H. van Jaarsveld2,3*, Bas Ponsioen2,3*, Cheryl Zimberlin2,4*, Ruben van Boxtel1,2, Arjan Buijs5, Norman Sachs1,2, Rene´ M. Overmeer2,3, G. Johan Offerhaus6, Harry Begthel1,2, Jeroen Korving1,2, Marc van de Wetering1,2,7, Gerald Schwank1,2, Meike Logtenberg1,2, Edwin Cuppen1,2, Hugo J. Snippert2,3, Jan Paul Medema2,4,
Geert J. P. L. Kops2,3 & Hans Clevers1,2

Crypt stem cells represent the cells of origin for intestinal neoplasia. Both mouse and human intestinal stem cells can be cultured in medium containing the stem-cell-niche factors WNT, R-spondin, epidermal growth factor (EGF) and noggin over long time periods as epithelial organoids that remain genetically and phenotypically stable. Here we utilize CRISPR/Cas9 technology for targeted gene modification of four of the most commonly mutated colorectal cancer genes (APC, P53 (also known as TP53), KRAS and SMAD4) in cultured human intestinal stem cells. Mutant organoids can be selected by removing individual growth factors from the culture medium. Quadruple mutants grow independently of all stem-cell-niche factors and tolerate the presence of the P53 stabilizer nutlin-3. Upon xenotransplantation into mice, quadruple mutants grow as tumours with features of invasive carcinoma. Finally, combined loss of APC and P53 is sufficient for the appearance of extensive aneuploidy, a hallmark of tumour progression.

The adenoma–carcinoma sequence proposes that the sequential acquisition of specific genetic alterations underlies the progression of colorectal cancer (CRC)1. Activation of the WNT pathway, most commonly through inactivating mutations in APC, initiates the formation of benign polyps. Progression is thought to occur through activating mutations in the EGF receptor (EGFR) pathway and inactivating mutations in the P53 and transforming growth factor (TGF)-b pathways2. Recent sequencing efforts have further explored the genomic landscape underlying CRC3. A major hurdle in iden- tifying essential driver mutations is that many CRCs have acquired either microsatellite instability or chromosomal instability (CIN), as tumours typically harbour hundreds to thousands of mutations. Using mouse models, Lgr51-intestinal stem cells were identified as cells of origin for intestinal neoplasia and were shown to fuel effective tumour growth4–6. A recent study has shown that dereg- ulation (by retroviral expression of short hairpin RNAs (shRNAs) or cDNA) of APC, P53, KRAS and SMAD4 is sufficient for trans- formation of cultured mouse colon into tumours with adenocarci- noma-like histology7. Of note, the reliance on paracrine growth factors provided by a mesenchymal component in this system does not allow a one-to-one correlation with the individual oncogenic mutations. Comparable human in vitro model systems to study tumour initiation and progression have not been developed. We have previously described ‘indefinite’ three-dimensional stem cell culture systems (organoids) derived from several organs including mouse and human small intestine, colon, pancreas and liver that remain genetically stable8–13.

Sequential introduction of CRC mutations
We set out to utilize CRISPR/Cas9-mediated genome editing14–16 to introduce four of the most frequent CRC mutations in human small intestinal organoid stem cell cultures. As the absolute knockout
efficiency is low, we made use of functional selection strategies to obtain clonal, mutant organoids. Since loss of APC is generally con- sidered to be an early event in CRC2, we first introduced inactivating mutations in APC. As previously described, withdrawal of WNT and R-spondin from the defined culture medium provides a functional selection for APC loss17 (Fig. 1a). Indeed, control-transfected orga- noids died when seeded in medium lacking WNT and R-spondin (Fig. 1b), whereas transfection of plasmids expressing Cas9 and single guide RNAs (sgRNAs) targeting APC in its mutation hotspot region allowed cystic clonal organoids to emerge (Fig. 1b, Extended Data Fig. 1a and Extended Data Table 1a). To obtain clonal cultures, indi- vidual organoids were expanded. Genotyping verified the presence of clonal insertions or deletions (indels) at the targeted regions (Extended Data Fig. 1b). Quantitative reverse transcription polymer- ase chain reaction (qRT–PCR) analysis for the WNT target gene AXIN2 confirmed the constitutive activity of the WNT pathway, as AXIN2 messenger RNA levels did not decrease upon WNT/R-spon- din withdrawal (Fig. 1c).
Next, we introduced inactivating mutations in P53 in APC knock- out (APCKO) intestinal organoids. We made use of nutlin-3 (ref. 18) to select for organoids with a functionally inactive P53 pathway (Fig. 1a). As expected, nutlin-3 stabilized P53 in intestinal organoids and acti- vated transcription of its target gene P21 (also known as CDKN1A) (Fig. 1e). Control sgRNA-transfected APCKO organoids died upon nutlin-3 treatment (Fig. 1d), whereas transfection of plasmids expres- sing Cas9 and sgRNAs targeting P53 enabled organoid outgrowth (Fig. 1d, Extended Data Fig. 1a and Extended Data Table 1a). Clonal expansion and genotyping verified the presence of frame- shift-inducing indels at the targeted loci (at the start of the DNA- binding domain, thereby yielding an inactive gene product; Extended Data Fig. 1c). Loss of P53 protein expression and P53 pathway inac- tivity were confirmed by western blot (Fig. 1e).

1Hubrecht Institute, Royal Netherlands Academy of Arts and Sciences (KNAW) and UMC Utrecht, 3584CT Utrecht, The Netherlands. 2Cancer Genomics Netherlands, UMC Utrecht, 3584CG Utrecht, The Netherlands. 3Molecular Cancer Research, Centre for Molecular Medicine, UMC Utrecht, 3584CG, Utrecht, The Netherlands. 4Laboratory of Experimental Oncology and Radiobiology, Centre for Experimental MolecularMedicine, AMC, 1105AZAmsterdam, The Netherlands. 5Department of Medical Genetics, UMC Utrecht, 3508AB Utrecht, The Netherlands. 6Departmentof Pathology, UMC Utrecht, 3584CX Utrecht, The Netherlands. 7Foundation Hubrecht Organoid Technology (HUB), 3584CT Utrecht, The Netherlands.
*These authors contributed equally to this work.

0 0 M O N T H 2 0 1 5 | V O L 0 0 0 | N A T U R E | 1

a
3 days 1–2 weeks
APCKO
Wild-type

(WENR medium)
APC sgRNA
Cas9
– Wnt – R-spondin

3 days 1–2 weeks
Pick single organoids for clonal outgrowth

APCKO/P53KO

APCKO (EN medium)

P53 sgRNA
Cas9

+ Nutlin-3

b d

Wild-type

sgRNA APC (1)

sgRNA APC (3)
sgRNA control

sgRNA APC (2)

sgRNA APC (4)
sgRNA control

sgRNA P53 (1)

sgRNA P53 (3)

sgRNA P53 (2)

sgRNA P53 (4)

0.03

WENR
EN

APC sgRNA3

APC sgRNA4

Wild-type

APC/P53 sgRNA
3 4

0.02

0.01

0
kDa
55
40
35
15
– + – + – + + + Nutlin-3 P53
GAPDH
P21

Wild-type sgRNA3 sgRNA4
APC

Figure 1 | Inactivation of APC and P53 in human intestinal organoids. a, Strategy to generate the indicated mutant lines using CRISPR/Cas9. Blue,
stem cells. E, EGF; N, noggin; R, R-spondin; W, WNT. b, Wild-type organoids in complete medium (WENR; top left) and transfected with Cas9 and the indicated sgRNAs selected in EN medium (representative pictures from n 5 3 independent experiments). c, qRT–PCR for AXIN2 in wild-type and APCKO organoids in the presence or absence of WNT/R-spondin. Expression
normalized to GAPDH. Horizontal bars represent mean of n 5 3 independent experiments. d, APCKO organoids were transfected with Cas9 and the indicated sgRNAs. P53 mutants were selected in medium with nutlin-3 (representative pictures from n 5 3 independent experiments). e, Western blot analysis of P53 and P21 expression in organoids cultured in the presence/absence of nutlin-3 (representative from n 5 3 independent experiments). Scale bars, 100 mm.

The most common KRAS mutation in CRC results in the express- ion of constitutively active KRAS(G12D). To introduce this mutation, we designed an oligonucleotide with the oncogenic mutation and two silent mutations to serve as a template for homologous recombination (Fig. 2a). KRASG12D mutants were selected by withdrawing EGF and adding the EGFR inhibitor gefitinib to the culture medium. Of note, resident Paneth cells produce EGF in organoids9. Whereas control- transfected organoids failed to expand in the selection medium, orga- noids transfected with the oligonucleotide, Cas9 and the KRAS sgRNA grew out (Fig. 2b and Extended Data Table 1a). Genotyping of clonally expanded organoids confirmed that the resistant clones harboured the KRASG12D mutation (Fig. 2c). The two silent mutations were also present in the recombined allele, verifying that the muta- tions were introduced using the provided template. Although the second KRAS allele did not recombine, it was often targeted by Cas9 endonuclease, resulting in a frameshift in the second allele.

Quadruple mutants do not need niche factors
We then set out to introduce combinations of CRC mutations. We used our KRAS(G12D)-expressing organoids to introduce inac- tivating mutations in APC, P53 and SMAD4 (Fig. 2d). To select for inactivating mutations in SMAD4, an essential downstream

component of the TGF-b and bone morphogenetic protein (BMP) pathways, we made use of the dependence of the intestinal organoids on the presence of the BMP pathway inhibitor noggin in the culture medium9. Using the described selection procedures, transfection of Cas9 with either APC or both APC and P53 sgRNAs yielded KRASG12D/APCKO and KRASG12D/APCKO/P53KO organoids, respect- ively (Fig. 2e, f and Extended Data Fig. 2a, b, e). Transfection of Cas9 together with sgRNAs targeting APC, P53 and SMAD4 yielded orga- noids growing in medium lacking EGF, WNT, R-spondin and noggin, to which nutlin-3 was added (Fig. 2e, Extended Data Fig. 2d and Extended Data Table 1a). Clonal expansion and sequencing of the targeted loci in APC and P53 verified frameshift-inducing indels. Sequencing of the targeted exon in SMAD4 in several different clones revealed a frameshift-inducing deletion in one allele and an in-frame deletion in the other allele (sgRNA 1, P356del; sgRNA 3, V370del) (Extended Data Fig. 2c). Western blot confirmed reduced protein expression in SMAD4-mutated organoids (Fig. 2f). As with all sgRNAs, SMAD4 sgRNAs target the mutation hotspot region, encod- ing the MH2 domain required for SMAD4 activity19,20. Recently, in- frame deletions of SMAD4 P356 and V370 were shown to occur in CRC21, indicating that in-frame indels at these locations yield an inactive gene product. Using a candidate off-target prediction tool,

2 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 5

a b

Human KRAS locus
Exon 1
sgRNA control
sgRNA KRAS

G12D
KRASG12D
gRNA

Targeting oligonucleotide Human wild-type intestinal organoids
*GGT>GAT*

Human KRASG12D
P2 P1

c
locus
* *GGT>GAT
d
5′-..CTGAATATAAACTTGTGGTAGTTGGAGCTGGTGGCGTAGGCAAGAGTGCCT..-3′ KRAS exon 1
5′-..CTGAATATAAACTTGTGGTAGTTGGAGCCGATGGAGTAGGCAAGAGTGCCT..-3′ Targeting
oligonucleotide
Recombined KRASG12D allele
KRASG12D
CTGAATATAAACTTGTGGTAGTTGGAGCCGATGGAGTAGGCAAGAGTGCCT
(WNR medium)

Frameshifted allele
CTGAATATAAACTTGTGGTAGTTGGAGCTGGTGGACGTAGGCAAGAGTGCCT

outgrowth
e Human KRAS(G12D) intestinal organoids

sgRNA control
sgRNA APC (4)

sgRNA APC/P53 (3)
sgRNA APC/P53/SMAD4 (1)

sgRNA APC/P53/SMAD4 (3)
sgRNA APC/P53/SMAD4 (2)

sgRNA APC/P53/SMAD4 (4)

kDa kDa

70
70
55
35
SMAD4 long SMAD4
P53
GAPDH
250
55
APC Tubulin

Figure 2 | KRASG12D/APCKO/P53KO/SMAD4KO organoids grow in the absence of stem-cell-niche factors in vitro. a, Strategy to introduce the KRASG12D mutation. Asterisks indicate silent mutations; arrowheads indicate genotype primers. b, Wild-type human intestinal organoids were transfected with Cas9, sgRNA and the oligonucleotide. KRASG12D mutants were selected in medium lacking EGF, with the EGFR inhibitor gefitinib (representative pictures from n 5 3 independent experiments). c, Sequence analysis of the targeted KRAS exon. Oncogenic GGT.GAT mutation is indicated in green;

silent mutations are in blue; protospacer adjacent motif (PAM) is underlined in red. d, Strategy to generate the indicated mutant lines using CRISPR/Cas9. Blue, stem cells. N, noggin (Nog); R, R-spondin; W, WNT. e, KRAS(G12D)- expressing organoids were transfected with Cas9 and the indicated sgRNAs (representative pictures from n 5 3 independent experiments). f, Western blot analysis of SMAD4, P53 and APC expression in the indicated organoid lines (representative from n 5 3 independent experiments). K, KRASG12D; A, APCKO; P, P53KO; S, SMAD4KO. Scale bars, 100 mm.

we detected no lesions of predicted off-target sites for the sgRNAs used to introduce mutation combinations in our human intestinal organoids (Supplementary Table 1). Although this analysis was lim- ited, in combination with the analysis of multiple independent clonal organoids, the results indicated that the observed effects were not due

tumours were larger, highly proliferative and all displayed features of invasive carcinoma, including an irregular multi-layered epithelium consisting of tumour cells with increased nuclear–cytoplasmic ratio, pleiomorphic and hyperchromatic nuclei. Invasion of isolated or small aggregates of tumour cells into the stroma was frequently

to off-target effects. In conclusion, KRASG12D
KO
/APCKO/P53
/
observed (Fig. 3b and Extended Data Fig. 4b, d). The tumour origin

SMAD4KO mutant intestinal organoids can grow in the absence of all stem-cell-niche factors in vitro (Extended Data Table 1b).

Quadruple mutants grow as invasive carcinomas
Next, we investigated whether our engineered organoids were tumori- genic in vivo. We subcutaneously injected wild-type and all engi- neered mutant organoid lines into immunodeficient mice. After 8 weeks, some mice injected with KRASG12D/APCKO/P53KO organoids (‘triple’; 3 out of 12 injections) and the majority of mice injected with KRASG12D/APCKO/P53KO/SMAD4KO organoids (‘quadruple’; 13 out of 16 injections) developed visible nodules (Extended Data Fig. 4a). Histological analysis confirmed that triple organoids did engraft, but remained small with few proliferating cells and mostly resembled adenomas (Fig. 3a and Extended Data Fig. 4b, c). Quadruple-derived
was verified using a human-specific cytokeratin antibody (hKRT; Fig. 3 and Extended Data Fig. 4c, d). Thus, introduction of oncogenic mutations in KRAS, APC, P53 and SMAD4 enables normal human intestinal stem cell organoids to grow as tumours with invasive car- cinoma features in vivo (Extended Data Table 1b). In vitro, both triple- and quadruple-mutant organoids exhibited a high prolifera- tion rate, while only quadruple-mutant organoids frequently appeared as solid tumour masses (Extended Data Fig. 5a, c).

Extensive aneuploidy upon APC and P53 loss
To determine whether our engineered mutant organoids acquired CIN, a hallmark of CRC22, we transduced all human intestinal orga- noid lines with a fluorescently tagged histone 2B (H2B)-encoding lentivirus. This enabled us to monitor chromosome segregations

0 0 M O N T H 2 0 1 5 | V O L 0 0 0 | N A T U R E | 3

a a
H&E hKRT
100

5–12 divisions 13–20 divisions
>20 divisions Provided as video

H&E

hKRT

Ki67
APC
P53
KRAS
SMAD4
WT
WT
WT
WT
KO
WT
WT
WT
KO
KO
WT
WT
KO
KO
G12D
WT
KO
KO
G12D
KO

b
–9′ t = 0 + 33′ + 45′ + 61′ + 99′

H&E hKRT
c

100

<44 50 0 >48
44–48

Wild-type
APC sgRNA3
APC sgRNA4
APC/P53 sgRNA3
APC/P53 sgRNA4
Triple Quadruple

H&E hKRT Ki67 d

H&E

hKRT

Ki67
Wild-type APCKO
Figure 4 | Progressive CIN and aneuploidy upon introduction of CRC mutations. a, Live-cell imaging was performed to monitor chromosome segregations. Graph shows the percentage of erroneous mitoses. Each dot represents the percentage of errors in one organoid. Horizontal bars represent median of all dots. Videos are included of organoids depicted as dots with green outline (Supplementary Videos 1–5). WT, wild-type; KO, knockout. b, Stills of
KO
a typical erroneous mitotic event (anaphase bridge) in an APCKO/P53 organoid. Time points are indicated in minutes relative to prophase onset. Scale

Figure 3 | Quadruple-mutant organoids grow as invasive carcinomas in vivo. a, Haematoxylin and eosin (H&E; top left, bottom left), hKRT (top right, bottom middle) and Ki67 (bottom right) immunostainings on nodules isolated from triple-mutant-injected mice. Representative pictures of an adenoma with regular glandular structures lined with a blander epithelium that only focally shows a tendency towards stratification (arrowhead), no invasive growth and low proliferative capacity (Ki67). n 5 3 mice. b, As in a, but for quadruple- mutant-injectedmice.Upper andlower boxed regionsin toppanelscorrespond to regions imaged in left bottom panels and middle and right bottom panels, respectively. Representative pictures of an invasive carcinoma with irregular glandular architecture, gland in gland formation (white arrowhead) and luminal debris (asterisk). Mitotic figures are encountered (arrows) and there is high proliferative activity (Ki67). Invasion of isolated or small aggregates of cells into the stroma is observed (black arrowheads). n 5 13 mice. Scale bars, 100 mm.

using three-dimensional live-cell imaging. Wild-type organoids underwent mitosis without showing any major abnormalities. We did not observe an increase in the percentage of errors in APCKO organoids (Fig. 4a and Supplementary Videos 1 and 2). However,
bars, 5 mm. c, Chromosomes were counted in the indicated organoids. Graphs plot the percentage of cells with chromosome counts ,44, 44–48 (normal) and
.48 (at least 50 spreads were counted). d, Representative karyotypes of a wild- type culture with n 5 46 chromosomal counts (left) and APCKO organoid culture (right) with aberrant chromosome numbers. Scale bars, 25 mm.

To verify that the observed CIN results in aneuploidy, we next counted chromosome numbers. Unlike wild-type organoids, karyo- typing reproducibly revealed numerical aberrations in a low percent- age of APCKO organoids (two independent sgRNAs) (Fig. 4c, d). This ranged from a trisomy of chromosome 7 to near-tetraploid meta- phases (Extended Data Fig. 6a), the latter confirming previous studies in mouse embryonic stem cells23. Strikingly, one of the most recurrent chromosomal aberrations in low-grade colorectal adenomas in patients involves copy number gains of chromosome 7 (refs 24–26). In accordance with the chromosome segregation analyses, APCKO/
P53KO organoids showed a marked increase in the percentage of aneuploid spreads (Fig. 4c and Extended Data Fig. 6b). The triple and quadruple mutants also showed extensive aneuploidy (Fig. 4c and Extended Data Fig. 6c, d).

APC
KO
/P53KO organoids showed a progressive increase in the per-
To confirm these data, APC
KO
, P53
KO
and APC
KO
/P53KO orga-

centage of errors. We mainly observed anaphase bridges, but a few misaligned and lagging chromosomes were also detected (Fig. 4a, b and
KO KO
Supplementary Video 3). Importantly, compared to APC /P53 organoids, triple and quadruple mutants showed only a minor increase in the percentage of mitotic errors (Fig. 4a and Supplementary Videos 3–5), implying that loss of APC and P53 is sufficient to acquire CIN.

4 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 5
noids were engineered in a second human small intestinal line (Extended Data Fig. 7a–c). Again, loss of both APC and P53 had the most dramatic effect on CIN and aneuploidy. Although the single loss of P53 resulted in a substantial increase in the percentage of segregation errors, only a minor increase in the amount of aberrant spreads was observed (Extended Data Fig. 7d, f and Supplementary Video 6). Thus, we show that the combined loss of APC and P53 is

sufficient for the appearance of extensive aneuploidy. Despite the observed chromosome missegregations, our engineered lines con- tinue proliferating, while maintaining functional DNA damage sig- nalling (Extended Data Fig. 7h).

CRC mutations in human colon organoids
Finally, we introduced all the mutation combinations described earl- ier into a human colon organoid stem cell culture27, following the same functional selection procedures (Extended Data Fig. 3 and Extended Data Table 1b). Importantly, this yielded essentially ident- ical results to those obtained with the small intestinal stem cells, in terms of growth factor independence, in vitro appearance, CIN and aneuploidy (Extended Data Figs 5b and 7e, g). Moreover, both triple- and quadruple-mutant human colon organoids grew with high effi- ciency as tumours upon xenotransplantation into immunodeficient mice (Extended Data Fig. 8a, b). Histological analysis revealed that triple-mutant tumours contained large cysts and locally displayed features of well-differentiated carcinomas with relatively limited invasive growth, whereas the quadruple-mutant-derived invasive car- cinomas were faster growing, had a poorly differentiated appearance and displayed very frequent tumour budding at the invasive front, as well as invasion of the underlying muscle tissue (Extended Data Fig. 8c–e).
While this manuscript was under final review, a study using a similar strategy appeared28. Our CRC progression model selects out functional mutants by changing the culture medium composition and all sgRNAs were designed to target mutation hotspot regions. Therefore, we believe that our model reflects the in vivo situation more closely than any other in vitro human CRC model so far. Upon oncogenic mutation of KRAS, APC, P53 and SMAD4, human gut stem cell organoids can grow in the absence of all stem-cell-niche factors and in the presence of the P53 stabilizer nutlin-3 in vitro and as tumours with invasive carcinoma features in vivo. Moreover, we find that our engineered CRC organoid lines show marked CIN and aneu- ploidy, both considered to be hallmarks of cancer22.
Online Content Methods, along with any additional Extended Data display items andSource Data, are available in the onlineversion of the paper; references unique to these sections appear only in the online paper.

Received 11 November 2014; accepted 16 March 2015. Published online 29 April 2015.
1.Fearon, E. R. & Vogelstein, B. A genetic model for colorectal tumorigenesis. Cell 61, 759–767 (1990).
2.Fearon, E.R.Molecular geneticsofcolorectalcancer. Annu. Rev. Pathol. 6, 479–507 (2011).
3.The Cancer Genome Atlas Network. Comprehensive molecular characterization of human colon and rectal cancer. Nature 487, 330–337 (2012).
4.Barker, N. et al. Crypt stem cells as the cells-of-origin of intestinal cancer. Nature 457, 608–611 (2009).
5.Schepers, A. G. et al. Lineage tracing reveals Lgr51 stem cell activity in mouse intestinal adenomas. Science 337, 730–735 (2012).
6.Zhu, L. et al. Prominin 1 marks intestinal stem cells that are susceptible to neoplastic transformation. Nature 457, 603–607 (2009).
7.Li, X. et al. Oncogenic transformation of diverse gastrointestinal tissues in primary organoid culture. Nature Med. 20, 769–777 (2014).
8.Sato, T. et al. Single Lgr5 stem cell build crypt-villus structures in vitro without a mesenchymal niche. Nature 459, 262–265 (2009).
9.Sato, T. et al. Long-term expansion of epithelial organoids from human colon, adenoma, adenocarcinoma, and Barrett’s epithelium. Gastroenterology 141, 1762–1772 (2011).
10.Sato, T. & Clevers, H. Primary mouse small intestinal epithelial cell cultures. Methods Mol. Biol. 945, 319–328 (2013).
11.Jung, P. et al. Isolation and in vitro expansion of human colonic stem cells. Nature Med. 17, 1225–1227 (2011).

12.Huch, M. et al. Unlimited in vitro expansion of adult bi-potent pancreas progenitors through the Lgr5/R-spondin axis. EMBO J. 32, 2708–2721 (2013).
13.Huch, M. et al. In vitro expansion of single Lgr51 liver stem cells induced by Wnt- driven regeneration. Nature 494, 247–250 (2013).
14.Mali, P. et al. RNA-guided human genome engineering via Cas9. Science 339, 823–826 (2013).
15.Cho, S. W., Kim, S., Kim, J. M. & Kim, J. S. Targeted genome engineering in human cells with the Cas9 RNA-guided endonuclease. Nature Biotechnol. 31, 230–232 (2013).
16.Cong, L. et al. Multiplex genome engineering using CRISPR/Cas systems. Science 339, 819–823 (2013).
17.Schwank, G. et al. Functionalrepair of CFTR by CRISPR/Cas9 inintestinalstem cell organoids of cystic fibrosis patients. Cell Stem Cell 13, 653–658 (2013).
18.Vassilev, L. T. et al. In vivo activation of the p53 pathway by small-molecule antagonists of MDM2. Science 303, 844–848 (2004).
19.Shi, Y., Hata, A., Lo, R. S., Massaque´, J. & Pavletich, N. P. A structural basis for mutational inactivation of the tumour suppressor Smad4. Nature 388, 87–93 (1997).
20.Chacko, B. M. et al. Structural basis of heteromeric smad protein assembly in TGF- b signaling. Mol. Cell 15, 813–823 (2004).
21.Fleming, N. I. et al. SMAD2, SMAD3 and SMAD4 mutations in colorectal cancer. Cancer Res. 73, 725–735 (2013).
22.Rajagopalan, H., Nowak, M. A., Vogelstein, B. & Lengauer, C. The significance of unstable chromosomes in colorectal cancer. Nature Rev. Cancer 3, 695–701 (2003).
23.Fodde, R. et al. Mutations in the APC tumour suppressor gene cause chromosomal instability. Nature Cell Biol. 3, 433–438 (2001).
24.Bomme, L. et al. Clonal karyotypic abnormalities in colorectal adenomas: clues to the early genetic events in the adenoma-carcinoma sequence. Genes Chromosom. Cancer 10, 190–196 (1994).
25.Herbergs, J. et al. Chromosome aberrations in adenomas of the colon. Proof of trisomy 7 in tumor cells by combined interphase cytogenetics and immunocytochemistry. Int. J. Cancer 57, 781–785 (1994).
26.Ried, T. et al. Comparative genomic hybridization reveals a specific pattern of chromosomal gains and losses during the genesis of colorectal tumors. Genes Chromosom. Cancer 15, 234–245 (1996).
27.Van de Wetering, M. et al. Prospective derivation of a Living Organoid Biobank of colorectal cancer patients. Cell http://dx.doi.org/10.1016/j.cell.2015.03.053 (in the press).
28.Matano, M. et al. Modelling colorectal cancer using CRISPR-Cas9-mediated engineering of human intestinal organoids. Nature Med. 21, 256–262 (2015).
29.Kim, K. A. et al. Mitogenic influence of human R-spondin1 on the intestinal epithelium. Science 309, 1256–1259 (2005).
30.Shaner, N. C. et al. A bright monomeric green fluorescent protein derived from Branchiostoma lanceolatum. Nature Methods 10, 407–409 (2013).
31.Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics 25, 1754–1760 (2009).
32.DePristo, M. A. et al. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nature Genet. 43, 491–498 (2011).
Supplementary Information is available in the online version of the paper. Acknowledgements We would like to thank H. M. Rodermond for help with in vivo
transplantation assays and members of the contributing laboratories for support. We thankA. Pronk, W.van Houdt andJ.van Gorpfor facilitating humancolon tissue. Weare grateful for support from the following: The Netherlands Organisation for Scientific Research (NWO-ZonMw) VENI grant to J.D. (91614138); University of Amsterdam (2012-5735) and The Dutch Digestive Diseases Foundation (MLDS) (FP13-07) to C.Z. and J.P.M.; Netherlands Institute of Regenerative Medicine (N.S. and G.S.); Dutch Cancer Society (KWF) (KWF/PF-HUBR 2007-3956 for H.B.; KWF Fellowship
UU2013-6070 for H.J.S.); Stand Up to Cancer/Stichting Vrienden van het Hubrecht (M.v.d.W.); NWO-ZonMw (116.005.002 for R.v.B.); and the CancerGenomics.nl (NWO Gravitation) program.
Author Contributions J.D. and H.C. conceived the project and wrote the manuscript. J.D. engineered and characterized all mutant organoid lines. R.H.v.J., B.P., H.J.S., R.M.O. and G.J.P.L.K. designed and performed live-cell imaging experiments. C.Z. and J.P.M. performed in vivo transplantation assays. R.v.B. and E.C. performed off-target analyses. J.D. performed karyotyping. A.B. made karyograms. G.J.O. staged subcutaneous tumours. H.B. and J.K. performed immunohistochemistry. N.S. optimized matrix for organoid growth. G.S. designed APC sgRNAs. M.v.d.W. established normal human colon organoid line. M.L. helped genotype the mutant small intestinal organoids.
Author Information Sequencing data have been deposited in the EMBL European Nucleotide Archive under accession number ERP009240. Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of the paper. Correspondence and requests for materials should be addressed to H.C. ([email protected]).

0 0 M O N T H 2 0 1 5 | V O L 0 0 0 | N A T U R E | 5

METHODS
Human material for organoid cultures. Approval for this study was obtained by the ethics committees of the University Medical Centre Utrecht (duodenal biops- ies) and The Diakonessen Hospital Utrecht (colonic tissues). Written informed consent was obtained.
Organoid culture. Endoscopic duodenal biopsy samples were obtained from two female individuals (patient 1, age 2 years; patient 2, age 8 years). These individuals were admitted for suspected coeliac disease or dyspepsia. Upon immunological and pathophysiological analysis, none of the individuals was diagnosed with coeliac disease, whereas patient 1 presented with signs of gastric metaplasia. All duodenal biopsies that were used in this study were found to be healthy on the basis of histological examination. Normal human colon tissue was isolated from a resected colon segment derived from a patient (female, age 60 years) diagnosed with CRC (sigmoid). Culture establishment was described previously17,27. Culture medium contains advanced DMEM/F12 medium (Invitrogen) including B27 (Invitrogen), nicotinamide (Sigma-Aldrich), N-acetylcysteine (Sigma-Aldrich), noggin (Peprotech), R-spondin 1 (ref. 29), EGF (Peprotech), WNT conditioned media (50%, produced using stably transfected L cells), TGF-b type I receptor inhibitor A83-01 (Tocris) and P38 inhibitor SB202190 (Sigma-Aldrich). For selection of KRASG12D mutants, organoids were grown in culture medium lacking EGF and containing 0.5–1.0 mM of gefitinib (Selleck Chemicals). For mutant P53 selection, organoids were cultured in the presence of 5–10 mM nutlin-3 (Cayman Chemical). Organoids were repeatedly tested for mycoplasma contamination and resulted negative.
Organoid transfection and genotyping. The organoid lipofection protocol was previously described in detail17. In short, human organoids were grown in the media described earlier, and trypsinized for 10 min at 37 uC. After trypsinization, cells were resuspended in 450 ml growth medium (containing the Rho kinase inhibitor Y-27632) and plated in 48-well plates at high density (80–90% conflu- ent). Nucleic acid–Lipofectamine 2000 complexes were prepared according to the standard Lipofectamine 2000 protocol (Invitrogen). Four microlitres of Lipofectamine 2000 reagent in 50 ml Opti-MEM medium (Gibco), and a total of 1.5 mg of DNA (sgRNA, Cas9, with/without oligonucleotide in 50 ml Opti- MEM medium) were mixed together, incubated for 5 min, and added to the cells (50 ml per well). The plate was centrifuged at 600g at 32 uC for 1 h, and incubated for 4 h at 37 uC before single cells were plated in Basement Membrane Extract (BME; Amsbio) or Matrigel (BD Biosciences). Growth medium plus Y-27632 was exchanged with selection medium 3 days after transfection. For clonal expansion single organoids were picked. On average, the efficiency of introduction of frame- shift-inducing mutations was approximately 1%. sgRNA transfections and sub- sequent selections were performed at least three times in both human small intestine and colon lines.
For genotyping, genomic DNA was isolated using Viagen Direct PCR (Viagen). Primers for the PCR amplification using GoTaq Flexi DNA polymerase (Promega) were as follows: APC_for, 59-TGTAATCAGACGACACAGGAAG CAGA-39, APC_rev, 59-TGGACCCTCTGAACTGCAGCAT-39; P53_for, 59- CAGGAAGCCAAAGGGTGAAGA-39, P53_rev, 59-CCCATCTACAGTCCCC

Western blot. Samples were lysed using RIPA buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 0.1% SDS, 0.5% Na-Deoxycholate, 1% NP-40) containing Complete protease inhibitors (Roche). Protein content was quantified using standard Bradford assay (BioRad) and equal amounts of protein were run on SDS–PAGE gels and transferred to PVDF membranes (Millipore). For APC western blotting, protein lysates were loaded on gradient polyacrylamide gels (4–15%; BioRad) and subsequently transferred. Membranes were blocked and probed with antibodies directed against P53 (DO-1, Santa Cruz Biotechnology), P21 (F-5, Santa Cruz Biotechnology), APC (FE9, Calbiochem), SMAD4 (B8, Santa Cruz Biotechnology), phospho-Chk1 Ser 345 (Bioke) and GAPDH (ab- 9485, Abcam). Organoid treatments: nutlin-3 10 mM, 24 h; doxorubicin 10 mM, overnight. Uncropped versions of the most relevant images are provided in Supplementary Fig. 1.
In vivo transplantation assays. Approval for this study was obtained by the Animal Experimentation Committee at the Academic Medical Centre in Amsterdam (DEC102581). Human organoid lines were expanded in their cor- responding selection media and trypsinized for 10 min at 37 uC. After trypsiniza- tion, 200,000 cells were resuspended in 50 ml of medium containing 23 required growth factors, mixed with Matrigel (BD Biosciences) at a 1:1 ratio and injected
tm1Wjl
subcutaneously into NOD scid gamma (NSG; NOD.Cg-Prkdcscid Il2rg /SzJ) mice ($6 injections per organoid line). After 7 (colon) or 8 (small intestine) weeks, mice were killed and nodules were processed for analysis. Both males and females (aged 8–10 weeks at the start of the experiment; weights, ,30 g for males and ,25 g for females) were used. This was randomly distributed and does not affect outgrowth. All animals were included in the analysis. Ear clipping was used for animal recognition. Number of injections was chosen following previous experience in the assessment of experimental variability. Animals were caged together and treated in the same way. Immunohistochemistry. Tissues were fixed in 4% paraformaldehyde, dehy- drated and embedded in paraffin. Sections were subjected to H&E as well as immunohistochemical staining. The following primary antibodies were used for immunohistochemical staining: anti-cytokeratin clone Cam5.2 (BD Biosciences), anti-Ki67 clone MM1 (Sanbio) and E-cadherin clone 36 (BD Biosciences).
Live-cell imaging and karyotyping. To visualize mitoses, organoids were infected with lentivirus encoding mNeon-tagged histone 2B and a puromycin- resistance cassette (pLV-H2B-mNeon-ires-Puro)30. After two passages, these were plated in BME in glass-bottom 96-well plates and mounted on an inverted confocal laser scanning microscope (Leica SP8X), which was continuously held at 37 uC and equipped with a culture chamber for overflow of 6.0% CO2. Over 16–20 h, ,10 H2B-mNeon-expressing organoids were imaged simulta- neously in XYZT-mode using a 340 objective (N.A. 1.1), using minimal amounts of 506 nm laser excitation light from a tuneable white light laser. Time interval was approximately 3 min (2:30–3:20 min). For post-acquisition analyses of mitotic behaviour, data sets were converted into manageable and maximally informative videos, combining z-projection, depth colour-coding and merging with transmitted light images (Supplementary Videos 1–6). Mitoses were scored, judged and counted manually.

CTTG-39; KRAS_for, 59-TGGACCCTGACATACTCCCA-39, KRAS_rev, 59-
21
For karyotyping, organoids were treated with 0.1 mg ml
colcemid (Gibco) for

AAGCGTCGATGGAGGAGTTT-39; SMAD4_for, 59-TGGAGTGCAAGTGA AAGCCT-39, SMAD4_rev, 59-ACCGACAATTAAGATGGAGTGCT-39. Prod- ucts were cloned into pGEM-T Easy vector system I (Promega) and subsequently sequenced using T7 sequencing primer.
Vector construction. The human codon-optimized Cas9 expression plasmid was obtained from Addgene (41815). The sgRNA-GFP plasmid was obtained from Addgene (41819) and used as a template for generating target-specific sgRNAs. The GFP targeting sequence was exchanged by inverse PCR followed by DpnI digestion and T4 ligation as described previously17. APC, P53 and SMAD4 sgRNA sequences are included in Extended Data Figs 1a and 2d. KRAS target sequences: number 1, 59-GAATATAAACTTGTGGTAGTTGG-39; number 2, 59- GTAGTTGGAGCTGGTGGCGTAGG-39.
RNA isolation, cDNA preparation and qRT–PCR. Organoids were harvested in RLT lysis buffer and RNA was isolated using the Qiagen RNeasy kit (Qiagen) according to the manufacturer’s instructions. Extracted RNA was used as a tem- plate for cDNA production using GoScript reverse transcriptase (Promega) according to the manufacturer’s protocol. qRT–PCR was performed using IQ SYBR green mix (Bio-Rad) according to the manufacturer’s protocol. Results were calculated by using the DDCt method. Organoid treatments: WNT/R-spon- din withdrawal, 48 h; nutlin-3 10 mM, 24 h. Primer sequences: AXIN2_for, 59- AGCTTACATGAGTAATGGGG-39, AXIN2_rev, 59-AATTCCATCTACACTG CTGTC-39; P21_for, 59-TACCCTTGTGCCTCGCTCAG-39, P21_rev, 59- GAGAAGATCAGCCGGCG TTT-39; GAPDH_for, 59-TGCACCACCAACTG CTTAGC-39, GAPDH_rev, 59-GGCATGGACTGTGGTCATGAG-39.
16 h. Cultures were washed and dissociated into single cells using TrypLE (Gibco) and processed as described13. Slides were mounted with DAPI-containing vecta- shield and analysed on a DM6000 Leica microscope (at least 50 spreads were analysed, n 5 3).
Off-target effect analysis. To assess off-target mutational effects, we computa- tionally identified candidate off-target sites for each sgRNA using COD software (http://cas9.wicp.net/). The software calculates an off-target score depending on sequence similarity: if the sequence perfectly matches the tested sgRNA (the target site) the score is 1 and decreases with increasing sequence differences. For the sgRNA targeting P53 and SMAD4 we identified 2 and 11 candidate off- target sites (Supplementary Table 1). For the sgRNA targeting APC and KRAS we only considered sites with an off-target scores of at least 0.15 or higher, resulting in 74 and 15 candidate off-target sites, respectively. We evaluated off-target mutational effects by amplicon-based NGS sequencing 93 candidate off-target sites and included the target sites for P53, KRAS and SMAD4 as positive controls (Supplementary Table 1). To this end, primers were designed ,350 nucleotides 59 and ,150 nucleotides 39 from the candidate site to obtain amplicons of ,500 bp (primer sequences available upon request). These regions were PCR amplified for each of the cultures using 5 ng genomic DNA, 13 GoTaq PCR Buffer (Promega), 1.5 mM MgCl2, 0.2 mM of each dNTP, 0.2 mM of each primer of a primer pair and 0.25 units of GoTaq polymerase (Promega) in a final volume of 10 ml at 94 uC for 60 s; 15 cycles at 92 uC for 30 s, 65 uC for 30 s with a decrement of 0.2 uC per cycle and 72 uC for 60 s; followed by 30 cycles at 92 uC for 30 s, 58 uC for 30 s and 72 uC for 60 s; and a final extension at 72 uC for 180 s. Per culture the PCR

products were pooled and barcoded. Illumina sequence libraries were generated using the TruSeq DNA Sample Preparation Kit (Illumina) according to the man- ufacturer’s protocol. Subsequently, the libraries were pooled and sequenced using the MiSeq sequencer (2 3 250 bp) to a depth of .10,0003 base coverage. Sequence reads were mapped to the human reference genome (GRCh37/hg19),

using the Burrows–Wheeler Aligner (BWA) Maximal Exact Matches (MEM) v.0.7.5a mapping tool31 with settings ‘-c 100 -m’. Small indel calling was per- formed using the Genome Analysis Toolkit (GATK)32 haplotype caller v.3.2-2 with ‘best practices’ settings. We only considered indels with a variant allele frequency (VAF) of at least 0.15 or higher (Supplementary Table 1).

Extended Data Figure 1 | Introducing inactivating mutations in the APC and P53 genes in human intestinal organoids using CRISPR/Cas9.
a, Schematic representation of the targeted exon of the human APC (left) and P53 (right) loci and sequences of the designed sgRNAs. b, c, PCR amplification products of the mutated alleles of APC (b) and P53 (c) were obtained using

primers flanking the targeted exon. Subsequent sequencing revealed indels at the expected locations. PAM sequences are underlined in red in wild-type sequences. Of note, the curved lines bridging the gaps in deleted alleles are drawn by the alignment software.

Extended Data Figure 2 | KRASG12D, APCKO, P53KO and SMAD4KO mutation combinations in human intestinal organoids. a–c, PCR amplification products of the indicated genes of KRASG12D/APCKO
(a), KRASG12D/APCKO/P53KO (b) and KRASG12D/APCKO/P53KO/SMAD4KO (c) organoids were obtained using primers flanking the targeted exon. Subsequent sequencing revealed indels at the expected locations. PAM sequences are underlined in red. Of note, the curved lines bridging the gaps in deleted alleles are drawn by the alignment software. d, Schematic

representation of the targeted exon of the human SMAD4 locus and sequences of the designed sgRNAs. e, qRT–PCR for AXIN2 (top) and P21 (bottom) in the indicated organoid cultures. Top, the indicated organoid lines were cultured in the presence (WENR) or absence (EN) of WNT/R-spondin. Bottom, the indicatedorganoid lines were culturedin thepresenceor absence ofnutlin-3for 24h. Expression wasnormalizedto GAPDH. Horizontalbars representmean of n 5 3 independent experiments.

Extended Data Figure 3 | Using CRISPR/Cas9-mediated genome editing to introduce APC, P53, KRASG12D and SMAD4mutations in human colonic organoids. a–d, Using the strategies depicted in Figs 1a and 2a, d, APCKO,
KO G12D
APC /P53KO (a), KRASG12D (b), KRASG12D/APCKO, KRAS /APCKO/
P53KO, KRASG12D/APCKO/P53KO/SMAD4KO (c) and P53KO (d) mutant human colon organoids were generated. Experiment was performed at least three independent times for each mutation. e, qRT–PCR for AXIN2 in the indicated organoid lines cultured in the presence (WENR) or absence (EN) of WNT/R- spondin. Expression was normalized to GAPDH. Horizontal bars represent mean of n 5 3 independent experiments. f, Western blot analysis of P53 and

P21 expression in the indicated human colon organoid lines cultured in the presence or absence of nutlin-3. GAPDH, loading control. g, qRT–PCR for AXIN2 in the indicated organoid lines cultured in the presence (WENR) or absence (EN) of WNT/R-spondin. Expression was normalized to GAPDH. Horizontal bars represent mean of n 5 3 independent experiments. h, Western blot analysis of SMAD4 and P53 expression in the indicated human colon organoid lines. Please note that quadruple-mutant clone 1 contains SMAD4 frameshift-inducing indels in both alleles whereas clone 2 contains a frameshift-inducing indel in one and an in-frame deletion in the other allele (reduced SMAD4 expression). GAPDH, loading control. Scale bars, 100 mm.

Extended Data Figure 4 | Quadruple-mutant human intestinal organoids grow as tumours with features of invasive carcinoma in vivo. a, Wild-type and all engineered human intestinal organoid lines were injected subcutaneously in immunodeficient mice. Mice injected with KRASG12D/
KO
APC /P53KO (triple) and KRASG12D/APCKO/P53KO/SMAD4KO (quadruple) organoids developed visible nodules. b, Tumour sizes were examined 8 weeks after transplantation. c, d, H&E (top left, bottom left), hKRT (top right, bottom

middle) and Ki67 (bottom right) immunostainings on nodules isolated from triple- (c) and quadruple-mutant (d) injected mice. Triple-mutant organoids did engraft but remained small, showed only weak proliferation and had adenoma features (n 5 3 mice). Quadruple-mutant-derived tumours were highlyproliferative withfeatures of invasive carcinoma(n 5 13 mice). See Fig. 3 for more details. Scale bars, 100 mm.

Extended Data Figure 5 | Histological analysis of triple- and quadruple- mutant organoids reveals morphological changes in vitro. a, Representative H&E and Ki67 immunostainings on the indicated human small intestinal organoid lines (n 5 4 independent experiments). b, Representative H&E
and Ki67 immunostainings on the indicated human colon organoid lines

(n 5 3 independent experiments). c, Representative E-cadherin immunostainings on wild-type and quadruple-mutant human small intestinal organoids (n 5 4 independent experiments). Asterisk indicates residual Matrigel. Scale bars, 100 mm.

Extended Data Figure 6 | Progressive aneuploidy upon introduction of
KO
CRC mutations. a–d, Karyograms of APCKO (a), APCKO/P53
(b), KRASG12D/APCKO/P53KO (c) and KRASG12D/APCKO/P53KO/SMAD4KO (d) organoids, showing extensive aneuploidy in organoids harbouring CRC

mutations (20 spreads were analysed per line). Note the occurrence of trisomy 7 in APCKO and APCKO/P53KO (independent clones) organoids. M, marker chromosomes.

Extended Data Figure 7 | Loss of both APC and P53 results in extensive CIN and aneuploidy. a, APCKO, P53KO and APCKO/P53KO mutations were introduced in a second independent human intestinal organoid line. PCR amplification products of the mutated alleles of APC and P53 were obtained using primers flanking the targeted exon. Subsequent sequencing revealed frameshift-inducing indels at the expected locations. Left, APC genotyping; right, P53 genotyping. PAM sequences are underlined in red. Of note, the curved lines bridging the gaps in deleted alleles are drawn by the alignment software. b, Western blot analysis for P53 and P21 expression in the second human intestinal organoid line cultured in the presence or absence of nutlin-3. GAPDH, loading control. c, qRT–PCR for AXIN2 in the second human intestinal organoid line cultured in the presence (WENR) or absence (EN) of WNT/R-spondin. Expression was normalized to GAPDH. Horizontal bar

represents mean of n 5 3 independent experiments. d, Chromosome numbers were counted in the second human intestinal organoid lines. Graphs plot the percentage of cells with chromosome counts ,44, 44–48 (normal) and .48 (at least 50 spreads were counted). e, As in d, but for indicated human colon organoid lines. f, Live-cell imaging was performed to monitor chromosome segregations in the indicated human small intestinal organoid lines. Graph shows the percentage of erroneous mitoses. Each dot represents the percentage of errors in one organoid. Horizontal bars represent median of all dots. A video is included of organoids depicted as dots with green outline (Supplementary Video 6). WT, wild type; KO, knockout. g, As in f, but for indicated human colon organoid lines. h, Western blot analysis of phospho-CHK1 and P53 expression in the indicated organoid lines treated with the DNA-damaging drug doxorubicin, or left untreated. GAPDH, loading control.

Extended Data Figure 8 | Engineered mutant human colon organoids grow as invasive carcinomas in vivo. a, Wild-type, triple- and quadruple-mutant human colon organoids were injected subcutaneously in immunodeficient mice. Nodules were counted 7 weeks after transplantation. b, Tumour sizes were examined 7 weeks after transplantation. c, Representative pictures of a ‘cystic’ triple-mutant (left) and ‘solid’ quadruple-mutant (right) tumour in immunodeficient mice. d, H&E (top left, bottom left), hKRT (top middle, bottom middle) and Ki67 (top right, bottom right) immunostainings on nodules isolated from triple-mutant-injected mice. Representative pictures of a

well-differentiated carcinoma with limited invasive growth. The invasive growth has an expansive growth pattern with little tumour budding. n 5 6 mice. e, As in d but for quadruple-mutant-derived tumours. Representative pictures of a poorly differentiated invasive carcinoma with frequent tumour budding at the invasive front (invasion of isolated or small aggregates of cells into the stroma is frequently observed (black arrowheads)). Invasive character is confirmed by the invasive growth into the underlying muscle tissue (asterisk, muscle tissue). n 5 8 mice. Scale bars, 100 mm.

Extended Data Table 1 | Introducing oncogenic mutations in human intestinal organoids using CRISPR/Cas9

a, Overview of the number of functional sgRNAs and the selection strategy used. b, Overview of the engineered lines. N.D., not determined.