305
Tumor Formation and Regulation in Xenopus
cysts. Already four to fve weeks after injection, tadpoles
showed organization abnormalities in the intestine, and desmoid tumors became apparent around metamorphosis (Van
Nieuwenhuysen et al., 2015). Of note, by using targeted injection of the apc TALEN mRNAs in a single blastomere of an
eight-cell embryo, it was found that the occurrence of specif c
tumor phenotypes and survival rates highly depended on the
blastomere that was injected, further demonstrating the virtue of targeted injections. For example, desmoid tumors were
not observed when embryos were injected animal-dorsally
(blastomere assignation) (Van Nieuwenhuysen et al., 2015).
As a side note, these TALEN injections eventually led to the
generation of a heterozygote apc line.
21.4.3. CRISPR/CAS9 CANCER MODELING
IN XENOPUS TROPICALIS
TALENs can be extremely useful as genome editing reagents
and are by their design and molecular mode of action less
prone to off-target effects. However, CRISPR, although
being PAM sequence restricted, has several interesting features over TALENs, such as higher in vivo effectiveness,
ease of use, shorter time required for reagent assembly, and
higher multiplexing potential (Nemudryi et al., 2014). These
unique characteristics were translated into the generation of
the frst CRISPR/Cas9 cancer models in Xenopus tropicalis.
Shortly after the f rst Xenopus tropicalis apc TALEN
tumor model, the f rst Xenopus tropicalis CRISPR/Cas9
tumor model was generated, closely mimicking human retinoblastoma by the concomitant biallelic inactivation of rb1
and rbl1 (Naert et al., 2016). Interestingly, while in humans,
biallelic disruption of the RB1 gene is suffcient to induce
retinoblastoma (Lohmann, 1999), the additional inactivation of the retinoblastoma-like rbl1 gene was required for
retinoblastoma formation in Xenopus. In fact, Xenopus
tropicalis thereby parallels the situation in the mouse, where
the simultaneous inactivation of Rb1 and Rbl1/p107 is also
required ( Robanus-Maandag et al., 1998). Importantly, retinoblastoma induction in the rb1/rbl1 crispant model showed
very high penetrance (up to 73%) and short latency (median
time of 69 days). In addition to the prevalent retinoblastoma,
occasional appearance of a brain tumor type (probably pinealoblastoma), as well as choroid plexus neoplasms and small
cell lung cancer, was observed (Naert et al., 2016).
Apart from the apc heterozygote line, generated via
TALENs, two novel cancer-prone Xenopus tropicalis lines
(tp53 +/− and tp53 −/−) were recently generated using
CRISPR/Cas9 (Naert, Dimitrakopoulou, et al., 2020). These
animals resemble the human Li-Fraumeni syndrome (LFS),
with manifestation of hematologic malignancies and sarcomas in direct accordance with previously published Tp53
mutant mouse models (Donehower et al., 1992; Jacks et al.,
1994). While tp53 +/− Xenopus animals showed normal survival in the frst two years, tp53 −/− animals already demonstrated clear signs of disease at the age of one year, and
only 34% of the animals were alive after two years (Naert,
Dimitrakopoulou, et al., 2020). It is important to notice that
the time until morbidity was substantially longer than is
the case for homozygous Tp53 knockout mice (4.5 months)
(Donehower, 1996 ), which we believe may be related to the
longer life span of Xenopus compared to the mouse.
Given the involvement of TP53 mutations as a cooperative
event in the vast majority of human cancers, the availability
of tp53 mutant lines also offers opportunities for new applications for a range of additional cancer models in Xenopus. In the
cancer models obtained by simultaneous disruption of rb1 and
rbl1, it was found that additional tp53 disruption aggravated
tumor malignancy for the choroid plexus carcinomas and the
gliomas (Konukiewitz et al., 2017; Naert, Dimitrakopoulou, et
al., 2020). In a pilot multiplexing experiment combining four
sgRNAs (rb1, rbl1, pten, and tp53), it could be determined that
pten disruption further aggrevated the glioblastoma phenotype
(Naert, Dimitrakopoulou, et al., 2020), which is in line with
clinical data (Xiao et al., 2002; Chow et al., 2011).
21.5. APPLICATION POTENTIAL OF XENOPUS
TROPICALIS CANCER MODELING
21.5.1. GEXM FOR IDENTIFICATION OF
NOVEL TUMOR DEPENDENCIES
Currently, there is an increased interest in the identif cation
of druggable gene products that are crucial for tumor growth.
These targets, further called dependency factors, are nowadays being identifed employing so-called “negative selection dependency screens” using techniques such as RNAi,
shRNA, and CRISPR/Cas9 (Chen et al., 2015, 2018; Tzelepis
et al., 2016; Tsherniak et al., 2017). Interestingly, recently,
a method for in vivo CRISPR/Cas9 Selection-mediated
Identifcation of Dependencies (CRISPR-SID) was developed using the apc -based Xenopus tropicalis desmoid tumor
model (Naert et al., 2021). Using a multiplexing strategy, a
guide RNA for the tumor suppressor gene apc is co-injected
with a guide RNA against a putative dependency gene, and
animals are grown until they develop desmoid tumors. Due
to their pure monoclonal and non-metastasizing growth, the
desmoid tumors are ideally suited for performing these targeted dependency screening approaches. Genotyping of the
isolated desmoid tumors allows the identifcation of dependency genes. More specifcally, one can quantitatively assess
selection mechanisms by comparing for a particular guide
RNA, the induced double strand break (DSB) repair outcomes in normal embryonic tissue in the absence of negative selection (determined by deep amplicon sequencing)
with the repair outcomes observed in experimental tumors.
Specifcally, the tumor will select for out-of-frame insertiondeletion (INDEL) mutations in tumor suppressor genes (=
positive selection) while favoring absence of gene editing
or the presence of in-frame INDELs in essential genes or
dependency genes. Hence, for a genuine dependency gene, in
all the tumors sampled, at least one of the two alleles of the
targeted dependency gene will be either wild type or have an
in-frame mutation that maintains protein function to allow
sustained tumor growth (Figure 21.2).
Tumor Formation and Regulation in Xenopus
cysts. Already four to fve weeks after injection, tadpoles
showed organization abnormalities in the intestine, and desmoid tumors became apparent around metamorphosis (Van
Nieuwenhuysen et al., 2015). Of note, by using targeted injection of the apc TALEN mRNAs in a single blastomere of an
eight-cell embryo, it was found that the occurrence of specif c
tumor phenotypes and survival rates highly depended on the
blastomere that was injected, further demonstrating the virtue of targeted injections. For example, desmoid tumors were
not observed when embryos were injected animal-dorsally
(blastomere assignation) (Van Nieuwenhuysen et al., 2015).
As a side note, these TALEN injections eventually led to the
generation of a heterozygote apc line.
21.4.3. CRISPR/CAS9 CANCER MODELING
IN XENOPUS TROPICALIS
TALENs can be extremely useful as genome editing reagents
and are by their design and molecular mode of action less
prone to off-target effects. However, CRISPR, although
being PAM sequence restricted, has several interesting features over TALENs, such as higher in vivo effectiveness,
ease of use, shorter time required for reagent assembly, and
higher multiplexing potential (Nemudryi et al., 2014). These
unique characteristics were translated into the generation of
the frst CRISPR/Cas9 cancer models in Xenopus tropicalis.
Shortly after the f rst Xenopus tropicalis apc TALEN
tumor model, the f rst Xenopus tropicalis CRISPR/Cas9
tumor model was generated, closely mimicking human retinoblastoma by the concomitant biallelic inactivation of rb1
and rbl1 (Naert et al., 2016). Interestingly, while in humans,
biallelic disruption of the RB1 gene is suffcient to induce
retinoblastoma (Lohmann, 1999), the additional inactivation of the retinoblastoma-like rbl1 gene was required for
retinoblastoma formation in Xenopus. In fact, Xenopus
tropicalis thereby parallels the situation in the mouse, where
the simultaneous inactivation of Rb1 and Rbl1/p107 is also
required ( Robanus-Maandag et al., 1998). Importantly, retinoblastoma induction in the rb1/rbl1 crispant model showed
very high penetrance (up to 73%) and short latency (median
time of 69 days). In addition to the prevalent retinoblastoma,
occasional appearance of a brain tumor type (probably pinealoblastoma), as well as choroid plexus neoplasms and small
cell lung cancer, was observed (Naert et al., 2016).
Apart from the apc heterozygote line, generated via
TALENs, two novel cancer-prone Xenopus tropicalis lines
(tp53 +/− and tp53 −/−) were recently generated using
CRISPR/Cas9 (Naert, Dimitrakopoulou, et al., 2020). These
animals resemble the human Li-Fraumeni syndrome (LFS),
with manifestation of hematologic malignancies and sarcomas in direct accordance with previously published Tp53
mutant mouse models (Donehower et al., 1992; Jacks et al.,
1994). While tp53 +/− Xenopus animals showed normal survival in the frst two years, tp53 −/− animals already demonstrated clear signs of disease at the age of one year, and
only 34% of the animals were alive after two years (Naert,
Dimitrakopoulou, et al., 2020). It is important to notice that
the time until morbidity was substantially longer than is
the case for homozygous Tp53 knockout mice (4.5 months)
(Donehower, 1996 ), which we believe may be related to the
longer life span of Xenopus compared to the mouse.
Given the involvement of TP53 mutations as a cooperative
event in the vast majority of human cancers, the availability
of tp53 mutant lines also offers opportunities for new applications for a range of additional cancer models in Xenopus. In the
cancer models obtained by simultaneous disruption of rb1 and
rbl1, it was found that additional tp53 disruption aggravated
tumor malignancy for the choroid plexus carcinomas and the
gliomas (Konukiewitz et al., 2017; Naert, Dimitrakopoulou, et
al., 2020). In a pilot multiplexing experiment combining four
sgRNAs (rb1, rbl1, pten, and tp53), it could be determined that
pten disruption further aggrevated the glioblastoma phenotype
(Naert, Dimitrakopoulou, et al., 2020), which is in line with
clinical data (Xiao et al., 2002; Chow et al., 2011).
21.5. APPLICATION POTENTIAL OF XENOPUS
TROPICALIS CANCER MODELING
21.5.1. GEXM FOR IDENTIFICATION OF
NOVEL TUMOR DEPENDENCIES
Currently, there is an increased interest in the identif cation
of druggable gene products that are crucial for tumor growth.
These targets, further called dependency factors, are nowadays being identifed employing so-called “negative selection dependency screens” using techniques such as RNAi,
shRNA, and CRISPR/Cas9 (Chen et al., 2015, 2018; Tzelepis
et al., 2016; Tsherniak et al., 2017). Interestingly, recently,
a method for in vivo CRISPR/Cas9 Selection-mediated
Identifcation of Dependencies (CRISPR-SID) was developed using the apc -based Xenopus tropicalis desmoid tumor
model (Naert et al., 2021). Using a multiplexing strategy, a
guide RNA for the tumor suppressor gene apc is co-injected
with a guide RNA against a putative dependency gene, and
animals are grown until they develop desmoid tumors. Due
to their pure monoclonal and non-metastasizing growth, the
desmoid tumors are ideally suited for performing these targeted dependency screening approaches. Genotyping of the
isolated desmoid tumors allows the identifcation of dependency genes. More specifcally, one can quantitatively assess
selection mechanisms by comparing for a particular guide
RNA, the induced double strand break (DSB) repair outcomes in normal embryonic tissue in the absence of negative selection (determined by deep amplicon sequencing)
with the repair outcomes observed in experimental tumors.
Specifcally, the tumor will select for out-of-frame insertiondeletion (INDEL) mutations in tumor suppressor genes (=
positive selection) while favoring absence of gene editing
or the presence of in-frame INDELs in essential genes or
dependency genes. Hence, for a genuine dependency gene, in
all the tumors sampled, at least one of the two alleles of the
targeted dependency gene will be either wild type or have an
in-frame mutation that maintains protein function to allow
sustained tumor growth (Figure 21.2).