DOI:10.1093/jnci/djt257
Advance Access publication September 19, 2013
© The Author 2013. Published by Oxford University Press. All rights reserved.
For Permissions, please e-mail: [email protected].
Review
Acquired Mutations That Affect Pre-mRNA Splicing
in Hematologic Malignancies and Solid Tumors
Linda M. Scott and Vivienne I. Rebel
Manuscript received April 9, 2013; revised June 5, 2013; accepted August 16, 2013.
Correspondence to: Linda M. Scott, PhD, University of Queensland, Translational Research Institute, 37 Kent Street, Woolloongabba, Queensland 4012,
Australia (email: [email protected]).
The application of next-generation sequencing technologies to interrogate the genome of human hematologic malignancies
is providing promising insights into their molecular etiology and into the pathogenesis of seemingly unrelated malignancies.
Among the somatic mutations identified by this approach are ones that target components of the spliceosome, a ribonucleoprotein complex responsible for the posttranscriptional processing of primary transcripts to form mature messenger RNA species.
These mutations were initially detected in patients with chronic lymphocytic leukemia or a myelodysplastic syndrome, but can
also occur at relatively high frequency in some solid tumors, including uveal malignant melanoma, adenocarcinoma of the lung,
and estrogen receptor–positive breast cancers. Their presence in a variety of malignancies suggests that the spliceosomal mutations may play a fundamental role in defining the malignant phenotype. The development and testing of drugs that eliminate cells
bearing a spliceosomal mutation, or normalize their altered transcript splicing patterns, are therefore a priority. Here, we summarize the effects of spliceosome-associated mutations on transcript processing in vitro and in vivo, and their impact on disease
initiation and/or progression and patient outcome. Moreover, we discuss the therapeutic potential of compounds already known
to target splicing factor 3B subunit 1 (SF3B1), an essential component of the spliceosome that is frequently mutated.
J Natl Cancer Inst;2013;105:1540–1549
The application of next-generation sequencing technologies to
interrogate the genome of human hematologic malignancies is
providing promising insights into their molecular etiology, and
into the pathogenesis of other seemingly unrelated malignancies.
Among the mutations identified by this approach are ones that
affect DNMT3A or IDH1 in patients with an acute myeloid leukemia (AML), or CREBBP in B-cell malignancies (1–4). Perhaps
the most interesting mutations detected thus far are those targeting constituents of the spliceosome, an RNA protein complex
responsible for the splicing process that transforms primary transcripts into mature mRNA species. Initially, these were detected
in patients with a myelodysplastic syndrome (MDS) (5–7), a heterogeneous grouping of hematologic malignancies characterized
by the presence of dysplastic changes that mostly occur within the
cytoplasmic granules or nucleus of blood cells, and by ineffective
hematopoiesis involving one or more myeloid lineages. Several
spliceosomal proteins mutated in MDS patients are also frequently
mutated in those diagnosed with chronic lymphocytic leukemia
(CLL), suggesting that this category of mutations may contribute
to the development of various malignancies. Indeed, within the last
year, whole genome sequencing has identified spliceosome-associated mutations in solid tumors, including those originating in the
breast, lung, and eye.
In this review, we summarize the spliceosome-associated molecular defects that have been identified to date, as well as their in vitro
and in vivo effects on transcript processing and their impact on
disease progression and patient outcome. Moreover, we discuss the
1540 Review | JNCI
therapeutic potential of compounds known to target SF3B1, one
of the frequently mutated spliceosome proteins. Finally, we raise
the possibility that some patients lacking spliceosomal mutations
may nonetheless have mutations that affect the splicing of nascent
transcripts.
The Process of mRNA Splicing
Posttranscriptional processing of primary transcripts to form
mature mRNA species involves the removal of intronic sequences
by the spliceosome, a complex that consists of five small nuclear
ribonucleoproteins (snRNPs) and between 100 and 300 associated
proteins. The “major” or U2-type spliceosome, which consists of
the U1, U2, U4, U5, and U6 snRNPs, catalyzes the vast majority of
transcript splicing events, whereas the “minor” or U12-type spliceosome utilizes the U12 snRNP to mediate splicing of approximately
800 specific transcripts. The complex multistep process that enables splicing has been comprehensively described in two excellent
review articles (8,9). For the purposes of this article, however, we
briefly summarize the steps involved in “major” transcript splicing. As shown in Figure 1, this process is initiated by the spliceosome assembly phase, which includes the sequential formation of
three complexes: an early or “E” complex, an “A” complex, and a
“B” complex. “E” complex formation results from the binding of a
U1 snRNP to the 5′ splice site, and the U2-associated factor, U2AF,
to the 3′ splice site and adjacent polypyrimidine tract. The U2AF
protein is a heterodimer consisting of a U2AF1 subunit (also called
Vol. 105, Issue 20 | October 16, 2013
Figure 1. Stepwise removal of intronic sequences from a transcript by
the spliceosome. Posttranscriptional processing of primary transcripts to
form mature mRNA species involves the removal of intronic sequences
by the spliceosome, a complex generally consisting of five small nuclear
ribonucleoproteins (snRNPs; U1, U2, U4, U5 and U6) and a large number of associated proteins. The splicing process is initiated by the binding
of U1 to the 5′ splice site, which includes a GU dinucleotide sequence,
and weak binding of the U2-associated factor (U2AF), to the 3′ splice site
(AG) and adjacent polypyrimidine (pY) tract, resulting in the formation of
an E complex. In the next phase, an ATP-dependent alteration in the transcript positioning brings into close proximity the two exons (labeled here
as I and II, respectively), facilitating binding of the U2 snRNP to the branchpoint sequence (denoted by #) to generate an A complex.The B complex is
formed upon the addition of a U4/U5/U6 snRNP triad to the spliceosome,
which then undergoes rearrangement to form the catalytically active C
complex, with loss of U2AF and the U1 and U4 snRNPs. Finally, the intervening intron is removed via two transesterification steps, so that exons
I and II lie immediately adjacent to one another in the mature transcript.
U2AF35) and a U2AF2 (or U2AF65) subunit, which mediate binding to the 3′ splice site and the polypyrimidine tract, respectively.
An ATP-dependent alteration in transcript positioning then brings
together the proximal and distal exons, facilitating the binding of
the U2 snRNP and generation of an “A” complex. This snRNP consists of a 12S RNA subunit and the SF3A and SF3B multiprotein
complexes, with SF3B1 mediating binding to the intronic branchpoint sequence. This motif serves as a docking site for the spliceosomal subunits essential for the transesterification steps involved in
splicing. The “B” complex is then formed by the addition of the
U4/U5/U6 snRNP triad, which enables spliceosome activation, in
which an ATP-dependent conformational rearrangement forms the
catalytically active “C” complex, with release of U2AF and the U1
and U4 snRNPs. Two transesterification steps result in excision of
the intron, and ligation of the exons such that the proximal and distal exons lie adjacent to one another in the mature transcript.
Most, if not all, of the spliceosomal proteins involved in the
removal of introns from the primary transcript also influence patterns of alternate splicing. Members of the serine/arginine (SR) family play a particularly crucial role in the regulation of this process
(10). The SR proteins contain one or more RNA binding domains
(RBDs) that mediate their ability to interact with intronic and exonic
splicing enhancers (ISEs and ESEs, respectively) present in many primary transcripts. Binding of SR proteins at these enhancer elements
enables recruitment of the U1 snRNP and U2AF to the neighboring
5′ and 3′ splice sites, respectively, promoting the inclusion of that
exon into the mature transcript. However, this process is regulated
by posttranslational modification; the methylation, acetylation, or
phosphorylation of SR proteins alters their ability to bind ISEs and
ESEs. Heterogeneous nuclear ribonucleoprotein complexes instead
bind to intronic or exonic splicing silencer elements (ISSs and ESSs),
blocking the recruitment of the U1 snRNP and U2AF by a mechanism that has not been fully defined. Most primary transcripts that
have been studied contain multiple ISE/ESE or ISS/ESS elements;
it is the overall balance between the activities at these two antagonistic classes of regulatory element that determines whether or not a
particular exon will be included in the mature transcript.
jnci.oxfordjournals.org
Defects in the Constituents of the
Spliceosome in Myeloid Malignancies
The first example of somatic mutations affecting components on
the spliceosome was reported in 2011, when whole exome sequence
analysis was performed on paired tumor and normal samples from
29 patients with MDS (5). Among this cohort, 16 individuals had
an acquired mutation that affected a component of the splicing
machinery (namely PRPF40B, SRSF2, SF3A1, SF3B1, U2AF1, or
ZRSR2). The mutation status of these six proteins, and another
three spliceosome-associated factors (SF1, SRSF1, and U2AF2),
was evaluated in an additional 582 samples from patients with a
myeloid malignancy. Mutually exclusive mutations were detected
in 209 cases, including in 6.6% of patients with de novo AML and
JNCI | Review 1541
9.4% of patients with a myeloproliferative neoplasm (MPN), and
occurred most frequently in the subset of MDS patients that are
positive for ringed sideroblasts, which are morphologically distinct erythroid precursor cells in which abnormal crystalline iron
deposits surround the mitochondria. This included 83% of patients
with refractory anemia with ringed sideroblasts (RARS), and 76%
of those with refractory cytopenia with multilineage dysplasia
and ringed sideroblasts (RCMD-RS). SRSF2 mutations were also
detected in more than a quarter of patients with chronic myelomonocytic leukemia, a clonal hematopoietic stem cell (HSC) disorder with characteristics of both MDS and an MPN. Subsequent
studies (6,7,11–15), several of which are summarized in Table 1,
have reported comparable incidences of spliceosome mutations in
adult patients with a myeloid malignancy. One of these studies also
showed that mutations may affect the LUC7L2, PRPF8, SF3B3,
SON, or SRSF6 spliceosome proteins (11). In contrast, spliceosomal mutations are relatively rare in cases of pediatric MDS
(occurring in 1 of 255 cases) (16), suggesting that the molecular
pathogenesis of juvenile and adult MDS are distinct.
The targets for the majority of spliceosome-associated mutations in MDS are components of the “E” and “A” splicing complexes
(Figure 1). U2AF, which is a heterodimer of the U2AF1 and U2AF2
proteins, binds to the 3′ splice site, bringing together adjacent exons
and facilitating the binding of the U2 snRNP. This snRNP includes
the SF3A and SF3B complexes; the former contains three protein subunits (SF3A1–3), and the latter consists of five (SF3B1–5).
ZRSR2 associates with the U2AF1 protein and plays a role in the
recognition the 3′ splice site, whereas SF1 recognizes the branchpoint sequence and is required for the generation of an “A” complex.
The SR proteins, including SON, SRSF1, SRFS2, and SRSF6, interact with other constituents of the spliceosome via their arginine/
serine-rich (RS) domains, forming a bridge that links together the
5′ and 3′ splice sites. LUC7L2, PRPF8, and PRPF40 are, however,
constituents of snRNPs other than U2. Relatively little is known
about the function of LUC7L2 or PRPF40B, which have homology to the yeast LUC7 and PRP40 splicing factors, respectively,
although LUC7L2 has a C2H2-type zinc finger and an RS domain
that suggest it is involved in mediating interactions between spliceosome components. Both proteins are present in the U1 snRNP,
whereas PRPF8 is a component of the U5 snRNP, where it interacts
with the 5′ and 3′ splice sites to align the proximal and distal exons
for subsequent ligation. PRPF8 is therefore essential for the second
transesterification step in pre-mRNA splicing.
The spliceosome mutations identified are predominantly simple
amino acid substitutions, although small in-frame deletions within
SRSF2 have been noted in patients with myelofibrosis or de novo
MDS (17,18). The location of the most frequently occurring alterations is shown in Figure 2. Most disease-associated mutations can
be localized to one of the functionally important domains in SF3B1,
U2AF1, or SRSF2. In SF3B1, most mutations map to one of the
22 HEAT Huntingtin, Elongation factor 3, protein phosphatase
2A, Targets of rapamycin 1 repeats present in this polypeptide,
and which mediate its interaction with other proteins. They cluster around residues 625, 666, and 700, with a K700E substitution
occurring most frequently. Mutations also commonly target either
of the zinc fingers present in U2AF1, with amino acid substitutions
occurring at residues S34 and Q157. In direct contrast, mutations
1542 Review | JNCI
do not affect the RBD or zinc finger of SRSF2, but instead target
residue P95, which lies between these domains. There are no mutational hotspots in ZRSR2, with alterations occurring throughout
the protein; several are amino acid substitutions that affect the RBD
or distal zinc finger motifs, although many are nonsense or frameshift mutations that result in protein truncation.
SF3B1 Mutations and Patient Outcome
in Patients Diagnosed With CLL
Acquired SF3B1 mutations also are present in a substantial proportion of patients with CLL (19–23) (Table 2). In contrast, the majority of the other spliceosome mutations associated with a myeloid
malignancy are conspicuously absent, although SRSF1, SRSF7,
or U2AF2 mutations can occur on occasion (22,23). Rare CLLassociated mutations may also target spliceosomal components that
are not affected in patients with MDS, including two that bind to
members of the SR family: SRPK1 and SRRM1 (23).
The CLL-associated SF3B1 mutations all localize to the HEAT
repeats, occurring in the same regions as the mutations present in
patients with MDS (Figure 2). Their frequency ranges substantially
between different cohorts (Table 2), which likely reflects both the
genetic diversity of patients within each cohort and their exposure
or response to chemotherapy. For example, the frequency of SF3B1
mutations in fludarabine-refractory patients was three times greater
than that of untreated patients (20). This difference is not solely
attributable to a more advanced disease stage in the former grouping, as patients who had undergone disease transformation to Richter
syndrome had a mutation frequency comparable to that of newly
diagnosed patients. SF3B1 mutations are more prevalent in patients
with a deletion of chromosome 11q22 or mutation in the ATM locus,
which lies within this region (22), and in patients without an IGHV
rearrangement (23). Accordingly, mutation-positive patients tend to
be diagnosed with a more advanced disease than those lacking an
SF3B1 mutation, and have a statistically significant lower 10-year
overall survival rate (23–25). However, the presence of an SF3B1
mutation did not appreciably alter overall survival or event-free survival rates in high-risk CLL patients receiving an allogeneic hematopoietic stem cell transplant (21). Analysis of the transcriptome of
patients with or without an SF3B1 mutation showed that the former express several alternatively spliced and prematurely truncated
transcripts that are not present in the latter grouping (23). Among
the most clinically relevant targets identified by this approach was
FOXP1, a transcription factor associated with the pathogenesis of
diffuse large B-cell lymphoma (26). A unique FOXP1 transcript that
is specific to CLL patients who have an SF3B1 mutation encodes a
shortened protein that lacks several of the PEST motifs that mediate its degradation (Figure 3A). This presumably alters the FOXP1
expression level in SF3B1 mutation–positive cells, thereby promoting proliferation, although this possibility remains to be tested.
Spliceosomal Mutations in
Nonhematologic Malignancies
Although the spliceosomal mutations were originally associated
with hematologic malignancies, recent next-generation sequencing studies have demonstrated that they may also occur in several
Vol. 105, Issue 20 | October 16, 2013
Table 1. Incidence of spliceosome-associated mutations in myeloid malignancies*
Incidence, %
Protein
Disease subtype
Yoshida (5)
PRPF40B
MDS w/o RS
RARS/RCMD-RS
CMML
MDS/sAML
De novo AML
MPN
MDS w/o RS
RARS/RCMD-RS
CMML
MDS/sAML
De novo AML
MPN
MDS w/o RS
RARS/RCMD-RS
CMML
MDS/sAML
De novo AML
MPN
MDS w/o RS
RARS/RCMD-RS
CMML
MDS/sAML
De novo AML
MPN
MDS w/o RS
RARS/RCMD-RS
CMML
MDS/sAML
De novo AML
MPN
MDS w/o RS
RARS/RCMD-RS
CMML
MDS/sAML
De novo AML
MPN
MDS w/o RS
RARS/RCMD-RS
CMML
MDS/sAML
De novo AML
MPN
MDS w/o RS
RARS/RCMD-RS
CMML
MDS/sAML
De novo AML
MPN
1.9
0.0
0.0
1.6
0.7
1.9
1.3
0.0
0.0
0.0
0.0
1.9
1.3
0.0
1.1
1.6
0.7
0.0
6.5
75.3
4.5
4.8
2.6
0.0
11.6
5.5
28.4
6.5
0.7
1.9
11.6
0.0
8.0
9.7
1.3
1.9
0.6
0.0
1.1
0.0
0.0
0.0
7.7
1.4
8.0
1.6
0.0
1.9
SF1
SF3A1
SF3B1
SRSF2
U2AF1
U2AF2
ZRSR2
Papaemmanuil (7)
7.0
64.6
4.7
ND
5.3
2.9
Makishima (11)
4.4
48.0
4.3
3.7
7.3
0.0
7.4
0.0
28.3
13.0
0.0
0.0
10.3
5.0
10.9
9.3
10.9
0.0
0.0
10.0
0.0
0.0
0.0
0.0
* AML = acute myeloid leukemia; CMML = chronic myelomonocytic leukemia; MDS = myelodysplastic syndrome; MPN = myeloproliferative neoplasm; ND = not
determined; PRPF40B = pre-mRNA processing factor 40 homolog B; RARS = refractory anemia with ringed sideroblasts; RCMD-RS = refractory cytopenia
with multi-lineage dysplasia and ringed sideroblasts; SF1 = splicing factor 1; SF3A1= splicing factor 3a, subunit 1; SRSF2 = serine/arginine-rich splicing factor 2;
U2AF1 = U2-associated factor, subunit 1; ZRSR2 = U2 small nuclear ribonucleoprotein auxiliary factor 35 kDa subunit-related protein 2.
solid tumor types, including neuroblastomas that arise following
chromothripsis (27), pancreatic ductal adenocarcinoma (28), and
estrogen receptor–positive tumors of the breast (29–31). Most
of these affect SF3B1, although mutations within SR proteins
(including SRSF1, SRSF12, and SRSF14) have been detected in
malignant breast tumors (31). Additional examples of spliceosomeassociated mutations occurring in solid tumors are to be expected
jnci.oxfordjournals.org
as the number of whole genome sequences reported in the literature rises. The true relevance of these spliceosome mutations is
also only just becoming apparent, with an unbiased analysis of the
genomes of 510 breast cancers identifying SF3B1 as one of the 35
most frequently mutated genes in this tumor type (32). Similarly, an
independent evaluation of 99 pancreatic tumors identified SF3B1
among the 16 most frequently mutated genes (28).
JNCI | Review 1543
Acquired heterozygous SF3B1 mutations were recently identified in a sizeable fraction (22 of 105) of patients with uveal malignant melanoma (33). In stark contrast to the mutation distribution
seen in patients with a hematologic malignancy, the melanomaassociated mutations target a single codon. This causes the substitution of arginine-625, located in the fifth HEAT repeat of SF3B1,
with a cysteine, glycine, histidine, or leucine. Within this cohort,
mutations in SF3B1 or BAP1, a nuclear ubiquitin carboxy-terminal hydrolase whose inactivation occurs in the majority of uveal
melanoma patients with a high metastatic risk (34), occurred in
a mutually exclusive fashion. Accordingly, the SF3B1 mutationpositive patients had a relatively favorable prognosis; compared to
mutation-negative patients, they presented at an earlier age, and
had tumors with a lower number of undifferentiated epithelial cells
(33). SF3B1 mutations do not feature prominently in primary or
metastatic melanomas originating in the hair-bearing skin (35), or
in patients with cutaneous melanoma (36).
Somatic U2AF1 mutations are also a recurring feature of lung
cancer (37), albeit with a lower frequency than the SF3B1 mutations in breast or pancreatic cancers. Overall, 10% of patients with
an adenocarcinoma of the lung were positive for a somatic mutation targeting a component of the splicing machinery. Four of the
183 tumor samples interrogated carried the U2AF1S34F mutation
originally described in MDS patients. Although the number of
mutation-positive cases was small, the presence of a U2AF1S34F
mutation was nevertheless associated with a reduction in the
progression-free survival rate. An additional patient had a G-to-I
substitution at residue 271 of U2AF1, which is located in the RS
domain; this variant has not been observed in patients with a hematologic malignancy. In an additional 14 cases, nonsynonymous
Figure 2. The nature and location of the most frequent spliceosome
mutations. The domain structure and mutation hotspots are depicted
for the spliceosome components most frequently mutated in human
malignancies. This includes SF3B1 (upper panel), U2AF1 (left side,
lower panel), ZRSR2 (middle, lower panel), and SRSF2 (right side,
lower panel). The SF3B1 HEAT repeats are represented as dark gray
lozenges, the zinc fingers in U2AF1 and ZRSR2 are white rectangles,
and the RNA-binding domain in U2AF1, SRSF2, and ZRSR2 are dark
gray rectangles. Diagonal lines highlight the arginine/serine-rich (RS)
domains of U2AF1, SRSF2, and ZRSR2. Most disease-associated mutations can be localized to a functionally important domain; those in
patients with MDS, CLL, or UM are indicated by triangles. An arrow
indicates the position of R1074, which confers in vitro resistance to
pladienolide B when mutated. In patients with MDS, ZRSR2 mutations
do not cluster together but are distributed throughout the protein; the
location of amino acid substitutions or premature truncations arising from nonsense or frame-shift mutations are indicated by asterisks. CLL = chronic lymphocytic leukemia; MDS = myelodysplastic
syndrome; SF3B1 = splicing factor 3B subunit 1; SRSF2 = serine/arginine-rich splicing factor 2; U2AF1 = U2-associated factor; UM = uveal
melanoma; ZRSR2 = U2 small nuclear ribonucleoprotein auxiliary factor 35 kDa subunit-related protein 2.
Table 2. Incidence of SF3B1 mutations in patients with chronic lymphocytic leukemia*
Study
(Ref)
CLL cohort particulars
Landau et al., 2013
Rossi et al., 2011
(19)
(20)
160 cases
301 untreated cases, 59 refractory to fludarabine, and
33 with Richter syndrome
Dreger et al., 2013
Wang et al., 2011
(21)
(22)
Quesada et al., 2012
(23)
100 high-risk cases
91 treated and untreated cases, including 22
with a del(11q)
279 untreated cases, including 76 with an IGHV
mutation and 73 lacking an IGHV mutation
Oscier et al., 2013
(24)
494 untreated patients recruited to the UK LRF CLL4 trial
Incidence
(%)
23 of 160
17 of 301
10 of 59
2 of 33
21 of 80
14 of 91
8 of 22
27 of 279
6 of 76
15 of 73
73 of 437
(14.4)
(5.6)
(16.9)
(6.1)
(26.3)
(15.4)
(36.4)
(9.7)
(7.9)
(20.5)
(16.7)
* CLL = chronic lymphocytic leukemia; IGHV = immunoglobulin heavy chain variable region.
1544 Review | JNCI
Vol. 105, Issue 20 | October 16, 2013
Figure 3. Spliceosome mutations may generate protein isoforms that
provide cells with a competitive advantage in vivo. A) The mutation
of SF3B1 in patients with chronic lymphocytic leukemia leads to the
expression of a truncated polypeptide, FOXP1w, which lacks several of
the PEST motifs that are present in FOXP1 (23). FOXP1w arises from
transcript splicing in tumor cells that does not occur in healthy cells:
a normally noncoding sequence that includes a termination codon
(denoted by an asterisk) is incorporated into the primary transcript
instead of exon 20, which encodes the PEST motifs and carboxy terminal domain. The abnormal transcript is not degraded by nonsensemediated decay; the truncated protein is thought to be stabilized by
the absence of PEST motifs, thereby promoting cell proliferation. B) In
healthy cells, the overall balance of alternate splicing patterns is governed by the balance between the activities of the splicing enhancer
elements (ISE/ESEs) and splicing silencer elements (ISS/ESSs).
Examples of normal differential splicing patterns are provided by the
MCL-1, BCL-X, and PKM loci. The full-length antiapoptotic MCL-1 protein, MCL-1L, is encoded by three exons (labeled here as I, II, and II).
A shorter isoform, MCL-1S, arises from a splicing event that excludes
exon II from the mature transcript. This alternate splicing replaces the
transmembrane domain with a nonsense domain (diamond motif),
which renders the protein proapoptotic. The BCL-X gene includes two
coding exons (A and B), with the former containing an internal 5′ splice
site (ISS). When this is utilized, a shorter mature transcript is produced;
this encodes the proapoptotic BCL-XS isoform. Splicing that includes
the distal portion of exon A (A′) in the mature transcript instead produces an antiapoptotic BCL-XL isoform. The use of alternative (mutually exclusive) exons may confer distinct properties to the resulting
protein: inclusion of exon 9 in the PKM transcript results in expression
of pyruvate kinase M1 isoenzyme (PKM1), which promotes oxidative
phosphorylation; exon 10 inclusion leads to PKM2 expression, promoting aerobic glycolysis. The balance in alternate splicing in malignant
cells may be perturbed by the mutation of components of the spliceosome, favoring the synthesis of isoforms that block apoptosis (MCL1L or BCL-XL) or provide a metabolic advantage by promoting a high
rate of glycolysis (PKM2). In other instances, missplicing may stabilize
protein expression levels (FOXP1), or produce truncated transcripts
that are degraded via nonsense-mediated mRNA decay. BCL-XL = long
isoform of BCL-X (B cell lymphoma-extra; BCL2L1); BCL-XS = short
isoform of BCL-X; FOXP1 = forkhead box protein P1; ISS = intronic
silencing splicer element; MCL-1 = induced myeloid leukemia cell differentiation protein; PKM-1, -2 = pyruvate kinase isozymes M1 and M2.
mutations were present in genes encoding other spliceosomal proteins, including U2AF2 (n = 4), SF3B1 or PRPF40B (n = 3), and
PRPF8, SF3A1, or SF3B3 (n = 1).
a reporter plasmid that contained three growth hormone exons and
two introns, substantially more exon skipping was detected in cells
that expressed U2AF1S34F (5,6). In agreement with this latter finding, the gene expression profiles of CD34-positive bone marrow cells
isolated from MDS patients with or without a U2AF1S34F mutation
(6), or with or without an SF3B1 mutation (7), showed a dramatic
change in the pattern of exon splicing. Among the genes upregulated in HeLa cells expressing mutant U2AF1, gene set enrichment
analysis also uncovered a statistically significant enrichment for genes
associated with nonsense-mediated mRNA decay (5), an intracellular
pathway that recognizes and degrades any abnormal transcripts with
the potential to prematurely terminate translation (as comprehensively reviewed in (38)). This abnormal expression pattern was normalized by U2AF1 overexpression, suggesting that U2AF1S34F may
be a dominant-negative protein that exerts its effects by inhibiting
the activity of its wild-type equivalent. Taken together, these studies suggest that the splicing activity of U2AF1 is both enhanced and
altered by its mutation, with affected cells activating the nonsensemediated decay pathway in response to the presence of abnormally
spliced transcripts. To date, studies have not been performed on the
functional consequences of SF3B1 mutations, so it remains unclear
whether all mutation variants have similar properties in vitro and
Functional Consequences of Spliceosome
Mutations In Vitro and In Vivo
Because of their nature, it was not immediately apparent whether the
spliceosome mutations identified in hematopoietic and solid tumors
enhanced or impaired pre-mRNA processing, or if they influenced
differential splicing patterns. However, the mutually exclusive nature
of the spliceosome mutations in patients with MDS suggests that
these may have similar effects in vivo, although this has not been
determined experimentally. However, insights into the consequence
of U2AF1 mutations were` provided by experiments testing the in
vitro activity of U2AF1S34F, the most common U2AF1 mutant. In
one study, a reporter plasmid that expresses luciferase only after an
intron containing translational stop codons is removed was introduced
into human kidney epithelial cells, together with wild-type or mutant
U2AF1 (6). Transcript splicing, as indicated by the luciferase activity, was notably higher in the presence of mutant U2AF1, suggesting that splicing was enhanced. In other experiments, which utilized
jnci.oxfordjournals.org
JNCI | Review 1545
in vivo. Therefore, additional studies are required to fully determine
the functional consequences of mutations targeting U2AF1, SF3B1,
and other spliceosome-associated proteins.
It is generally accepted that during tumorigenesis, cells of
the malignant clone evolve by accruing extra mutations that provide them with a competitive advantage over their sister cells.
Surprisingly, then, the presence of a U2AF1 mutation appears to
compromise the reconstituting ability of affected HSCs (5). In this
study, lethally irradiated mice were transplanted with bone marrow
cells transduced with retroviruses encoding wild-type or mutant
U2AF1; chimerism in the peripheral blood was assessed over time
using the coexpressed green fluorescent protein. Six weeks after
transplantation, the proportion of green fluorescent protein–positive cells in the recipients of donor cells expressing U2AF1Q157P,
U2AF1Q157R, or U2AF1S34F was less than that of recipients of
cells expressing wild-type U2AF1, suggesting that HSC engraftment, proliferation, or differentiation is impaired by mutant U2AF1.
However, these observations are in direct contrast to the finding that
the acquisition of an SF3B1 mutation in patients with CLL provides
mutation-bearing cells with a proliferative advantage (16). A similar
analysis in MDS patients may be necessary to determine whether
these mutations are advantageous in a myeloid cell context.
Two animal models of SF3B1 haploinsufficiency have provided
further insights into the consequences of an acquired SF3B1 mutation in patients with a hematologic malignancy. An N-ethyl-Nnitrosourea–based mutagenesis screen using zebrafish embryos
identified a lethal mutant (toastb460) that was characterized by cell
death in the dorsal central nervous system and by an absence of
circulating blood (39). These animals were found to have a hypomorphic Sf3b1 allele that encodes a truncated, nonfunctional Sf3b1
transcript. The expression level of several critical regulators of neural crest development was dramatically reduced in toastb460 embryos
as a result of defective pre-mRNA processing, which in turn caused
defects in lineage specification, cell survival, and migration within
the neural crest. The blood defect has not yet been assessed, although
it is plausible that defects in lineage specification, survival, and/or
migration may also be responsible for their hematologic phenotype.
Murine gene-targeting experiments revealed that, although Sf3b1null embryos die at the 32- to 64-cell stage, those heterozygous for
an Sf3b1 knockout allele remain viable and healthy (40). As adults,
these mice have skeletal anomalies, including a reduced number of
ribs, and mild lymphopenia (41). An elevated number of ringed sideroblasts was also noted when Sf3b1-haploinsufficient bone marrow
cells were stained with Prussian Blue, a dye that detects the crystalline iron deposits present within this cell type (41). Treatment of
healthy human bone marrow cells with meayamycin B, a drug that
affects mRNA splicing by inhibiting SF3B binding, similarly induced
the formation of ringed sideroblasts in vitro. These data collectively
suggest that the presence of ringed sideroblasts in individuals with
RARS or RCMD-RS may be directly attributable to a reduction in
SF3B1 activity within their erythroid precursor cells.
Therapeutic Potential of Drugs That Target
the Spliceosome
To identify compounds with therapeutic potential—that selectively
remove or normalize those cells that are positive for a spliceosome
1546 Review | JNCI
mutation—the effects that these mutations have on transcript processing will need to be more fully understood. Pharmacological
intervention might prove particularly challenging if these mutations encode inactive proteins, or proteins that have dominant-negative activity. If they provide a gain-of-function, however, splicing
patterns could potentially be normalized by exposure to one of the
already identified compounds that inhibit spliceosome function.
In a screen of natural compounds to identify those with antitumorigenic properties, three bacterial fermentation products
(FR901464, herboxidiene, and pladienolide) were identified as cytotoxic to multiple tumorigenic cell lines (42). Synthetic analogs with
improved stability and solubility, such as meayamycin, E7107, and
spliceostatin A (SSA), were subsequently developed. These drugs
all inhibit spliceosome function by specifically targeting the SF3B
protein complex, although it is unclear whether they target SF3B1,
SF3B3, or the interface between these subunits (43,44). Studies suggest that E7107 and SSA both destabilize U2 snRNP assembly at
3′ splice sites by impairing the ability of SF3B to bind RNA (42):
E7107 induces a conformational shift in the U2 snRNP that prevents its interaction with the branchpoint sequence, whereas SSA
induces it to bind to “decoy” sequences that can occur upstream of
the branchpoint sequence. However, it has been suggested that these
drugs do not globally inhibit splicing at therapeutic concentrations,
but instead perturb the expression of one or more proteins essential to the viability of a particular tumor cell type (42). An example
of this phenomenon was recently observed in two non–small cell
lung carcinoma cell lines that lack a spliceosome mutation (45).
Meayamycin B treatment of these cells induced cell death by altering the relative amounts of the short and long isoforms of MCL-1,
while not affecting the isoforms of another regulator of apoptosis,
BCL-X. Although meayamycin B acted as an overall inhibitor of
MCL-1 pre–mRNA processing, it predominantly altered the pattern of exons included in the mature transcript. This reduced the
relative level of the anti-apoptotic MCL-1L isoform, with a concomitant increase in the level of the proapoptotic MCL-1S isoform,
triggering cell death. This effect could be recapitulated by SF3B1
knockdown in these cell lines, confirming that meayamycin B treatment targeted the SF3B complex (45).
Although they result in tumor cell death in vitro, one concern is
that the SF3B inhibitors may mimic or even exacerbate the effects of
a disease-associated SF3B1 mutation. An example of this possibility
was outlined earlier in this text: normal erythroid precursors differentiate into ringed sideroblasts following exposure to meayamycin B
(41). The activity of these inhibitors will need to be tested on hematopoietic cells from patients with an SF3B1 mutation to clarify this
issue. Another concern is likely to be drug resistance: an R1074H
mutation in SF3B1 has already been identified and shown to confer
pladienolide B resistance to colorectal carcinoma cells in vitro (46).
Mutations at this residue have not been detected thus far in patients,
but nevertheless have the potential to arise and be selected for during treatment, impairing the in vivo efficacy of this compound. As
R1074 is located within the fifteenth HEAT repeat and the diseaseassociated SF3B1 mutations occur in other HEAT repeats (mainly
in repeats 5–8; Figure 2), cells carrying an SF3B1 mutation might
also have an impaired response to pladienolide B in vitro or in vivo.
An alternative strategy might be to employ natural regulators of spliceosomal activity to correct the effect of any mutation.
Vol. 105, Issue 20 | October 16, 2013
The SR proteins provide an ideal target, as several forms of posttranslational modification modulate their activity in healthy cells.
This includes arginine methylation, which influences subcellular
localization; acetylation of lysine residues, which may influence the
stability of SR proteins; and phosphorylation of serine residues in
the RS domain. Phosphorylation promotes spliceosome assembly
by inhibiting any random binding to RNA and by facilitating interactions between SR proteins, and also influences alternate splicing
patterns. SR proteins can be phosphorylated by topoisomerase I,
and members of the SRPK (SR protein kinase) and CLK (Cdc2like kinase) families. A number of small molecules are known to
inhibit SR protein phosphorylation, including two that are already
approved for the treatment of other conditions. Amiloride, an epithelial Na+ channel blocker used to treat hypertension, decreases
SRSF1 phosphorylation (35), and chlorhexidine, a topical antiseptic, inhibits the activity of CLK family members (21). Other
molecules that may have therapeutic potential include TG003, an
inhibitor of CLK1 and CLK4 (27); SRPIN340, an SRPK1 and
SRPK2 inhibitor (31); and NB-506, which inhibits the kinase activity of topoisomerase I (37).
A third potential approach involves altering the processing of key
pre-mRNAs by targeting the regulatory sequences (ISEs and ESEs,
or ISSs and ESSs) that modulate their differential splicing patterns.
Modified antisense oligonucleotides could be used to modulate the
function of these elements, increasing production of a beneficial
protein isoform or decreasing a deleterious isoform. This approach
has proved successful in a murine model of spinal muscular atrophy, a neurodegenerative disorder resulting from the inheritance of
mutations that reduce the level of functional survival motor neuron (SMN) protein (47). Blocking oligonucleotides were generated
against an ISS that prevents the inclusion of exon 7 in the mature
SMN2 transcript; treatment of affected mice with these oligonucleotides corrected SMN2 splicing, restoring SMN expression in motor
neuron cells and alleviating the disease phenotype. Phase I trials of
these oligonucleotides in patients with spinal muscular atrophy are
now under way. The synthesis of proteins with one or more isoforms
that support tumor cell proliferation, such as MCL-1, BCL-X, PKM
(Figure 3B), and VEGF-A, could be targeted in this fashion.
Are There Splicing Defects in Patients
Lacking the Spliceosome-Associated
Mutations?
Although a proportion of patients with MDS or CLL have acquired
mutations in components of the pathway responsible for primary
transcript processing, the majority of patients are unaffected. To
date, studies that have not used an unbiased whole genome or
exome sequencing strategy to detect spliceosome mutations, but
have rather employed a targeted sequencing approach, have not
looked at the entire repertoire of genes that encode spliceosome
proteins. It can therefore be anticipated that a subset of patients
who carry spliceosome mutations has not yet been identified. Do
some or all of the remaining patients have any abnormalities in
primary transcript processing? The detection of CPSF2, DDX3X,
and XPO1 mutations in patients with CLL who lack a spliceosomal mutation (22) support the hypothesis that perturbation of
mRNA processing may play an important role in the initiation
jnci.oxfordjournals.org
or progression of this disease, as CPSF2 participates in mRNA
polyadenylation, and DDX3X and XPO1 are both involved in
the nuclear export of transcripts (48–50). CLL patients may also
acquire mutations in EIF4A3 or MAGOH, two of the constituents of the exon junction complex (51). This complex forms at the
junction between adjacent exons following spliceosome-mediated
excision of the intervening sequence (Figure 1), and plays several
roles in posttranscriptional processing, including nuclear export.
Constituents of other complexes involved in RNA processing,
including the retention and splicing, cap-binding, and PRP19
complexes, may also be mutated (51).
Two recently published knockout mouse strains with a phenotype similar to human MDS suggest that patients with this disorder may also carry mutations that alter the correct processing of
biologically important transcripts. One of these studies involved
the deletion of Dicer, which encodes a microRNA (miRNA)
endonuclease, from the osteo­progenitor cells present within the
bone marrow microenvironment (52). This resulted in the emergence of an MDS-like phenotype, with mice developing anemia,
leukopenia, and thrombocytopenia, and marked dysplasia in
megakaryocytic cells. In contrast, loss of Dicer specifically from
HSCs increased their number by altering the proportion undergoing apoptosis (53). The cell-extrinsic nature of the defect in
Dicer-deficient osteoprogenitors was confirmed when the myelodysplastic phenotype was recapitulated by transplanting wildtype hematopoietic cells into Dicer-deficient recipients (52).
Inactivation of DICER may not contribute to MDS development
in humans, as DICER mutations have not been detected. However,
there is an intriguing connection between DICER deficiency and
Shwachman-Diamond-Bodian syndrome, a bone marrow failure syndrome that includes myelodysplasia among its features
and results from the inheritance of a mutant SBDS allele (54).
Gene set enrichment analysis of Dicer-deficient osteoprogenitor
cells revealed that Sbds transcription was dramatically reduced,
and targeted deletion of Sbds in these cells resulted in leukopenia
and dysplastic changes in megakaryocytes and granulocytes (52).
Acquired SBDS mutations do not occur in patients with de novo
MDS, although reduced expression of SBDS, as well as DICER
and DROSHA, another miRNA endonuclease, has been noted in
mesenchymal cells obtained from these individuals (55). The level
of each protein in mononuclear cells collected from these patients
and from healthy donors was comparable, however, suggesting
that unidentified factors in the marrow microenvironment may
contribute to the overall disease phenotype. Studies are required
to determine whether acquired mutations occur within the bone
marrow stroma of MDS patients, and if so, to identify what their
specific role is in the development of these disorders.
The second murine model linking abnormal splicing to the
development of MDS is the Crebbp heterozygous mouse (56–58).
It has long been appreciated that patients with therapy-related
MDS may carry a translocation between chromosomes 11 and
16, which fuses together the MLL and CREBBP genes (59). The
carboxy portion of CREBBP, which includes its bromodomain
and histone acetyltransferase domain, is retained in the resulting
chimeric protein, but the domain which binds nuclear hormone
receptors is invariably lost. Mll-Crebbp expression in a murine
transduction/transplantation model causes a myeloproliferative
JNCI | Review 1547
phenotype that undergoes eventual leukemic transformation (60).
In contrast, Crebbp-haploinsufficient mice initially develop a progressive pancytopenia (56). Reduced Crebbp levels do not influence HSC formation during embryogenesis nor perturb their
ability to differentiate in vitro, but affect their ability to self-renew
(57). Subsequent analyses revealed that the hematological disorder
in these mice has myelodysplastic and myeloproliferative features,
and is therefore an ideal animal model for studying human MDS/
MPN overlap syndromes (61). Whereas the MDS-like phenotype
is intrinsic to hematopoietic cells (V.I. Rebel, unpublished data),
the myeloproliferation is a consequence of reduced Crebbp levels
in the stromal and/or endothelial cells of the marrow microenvironment (58). Over time, about 40% of Crebbp-haploinsufficient
mice spontaneously develop an acute leukemia that is most often
myeloid in nature, with the loss of the remaining wild-type Crebbp
allele accompanying disease transformation (56).
Microarray-based analysis of the transcriptome of HSCs from
the fetal liver of Crebbp+/- mice and their wild-type littermates
revealed that there are major differences in gene expression in
these populations (62,63), as one might expect given the multiple
roles CREBBP plays in coordinating and regulating gene transcription. Strikingly, more than half of sequences downregulated
in Crebbp-haploinsufficient cells were intronic in origin, suggesting that they had smaller pools of unprocessed primary transcripts
(62). Among the genes affected were those that encode proteins
involved in RNA processing, including two SR proteins, Rbm39
and Sfrs12 (also referred to as Srek1); Prpf40A; Dgcr8, a doublestranded RNA binding protein that facilitates miRNA processing
by Drosha and Dicer (64); and Musashi-2, an RNA binding protein
whose overexpression alters HSC function in vivo and enhances
the leukemogenicity of the BCR/ABL oncoprotein (65). Another
downregulated target was Malat1, a large noncoding RNA whose
abundance is regulated by Dgcr8 and Drosha, and in turn regulates RNA splicing by modulating SR protein phosphorylation
(64). These data collectively suggest that mutations that decrease
CREBBP activity in hematopoietic cells substantially alter transcript processing within these cells, thereby contributing to the
development of MDS.
Conclusions
With the advent of DNA sequencing technologies that for the first
time allow rapid interrogation of the human genome, remarkable
progress has been made in our understanding of the molecular
pathogenesis of many malignancies. Of particular interest are the
acquired mutations that affect components of the human spliceosome. Clinically relevant mutations that alter transcript splicing
have been known for many years; estimates suggest that 15% of
disease-causing mutations are located in the 5′ or 3′ splice sites,
and that more than 20% of missense mutations lie in predicted
splicing enhancer or silencer elements (http://www.dbass.org.uk).
However, it was not appreciated until 2011 that disease-associated
mutations might also target components of the splicing machinery. The identification of these mutations not only in patients with
a hematological malignancy, but also in an expanding list of solid
tumors, suggests that they play a fundamental role in establishing
or maintaining the malignant phenotype. Drugs that normalize the
1548 Review | JNCI
changes in splicing associated with these mutations, or that target
for degradation those cells bearing a mutation, should therefore be
considered a priority.
References
1.Mardis ER, Ding L, Dooling DJ, et al. Recurring mutations found
by sequencing an acute myeloid leukemia genome. N Engl J Med.
2009;361:1058–1066.
2. Ley TJ, Ding L, Walter MJ, et al. DNMT3A mutations in acute myeloid
leukemia. N Engl J Med. 2010;363:2424–2433.
3. Mullighan CG, Zhang J, Kasper LH, et al. CREBBP mutations in relapsed
acute lymphoblastic leukaemia. Nature. 2011;471(7337):235–239.
4.Pasqualucci L, Dominguez-Sola D, Chiarenza A, et al. Inactivating
mutations of acetyltransferase genes in B-cell lymphoma. Nature.
2011;471(7337):189–195.
5.Yoshida K, Sanada M, Shiraishi Y, et al. Frequent pathway mutations of
splicing machinery in myelodysplasia. Nature. 2011;478(7367):64–69.
6. Graubert TA, Shen D, Ding L, et al. Recurrent mutations in the U2AF1
splicing factor in myelodysplastic syndromes. Nat Genet. 2011;44(1):53–57.
7.Papaemmanuil E, Cazzola M, Boultwood J, et al. Somatic SF3B1
mutation in myelodysplasia with ring sideroblasts. N Engl J Med.
2011;365(15):1384–1395.
8.Chen HC, Cheng SC. Functional roles of protein splicing factors. Biosci
Rep. 2012;32(4):345–359.
9. Wahl MC, Will CL, Luhrmann R. The spliceosome: design principles of a
dynamic RNP machine. Cell. 2009;136:701–718.
10.Zhou Z, Fu XD. Regulation of splicing by SR proteins and SR proteinspecific kinases. Chromosoma. 2013;122(3):191–207.
11.Makishima H, Visconte V, Sakaguchi H, et al. Mutations in the spliceosome machinery, a novel and ubiquitous pathway in leukemogenesis. Blood.
2012;119(14):3203–3210.
12. Thol F, Kade S, Schlarmann C, et al. Frequency and prognostic impact of
mutations in SRSF2, U2AF1, and ZRSR2 in patients with myelodysplastic
syndromes. Blood. 2012;119(15):3578–3584.
13.Visconte V, Makishima H, Jankowska A, et al. SF3B1, a splicing factor,
is frequently mutated in refractory anaemia with ringed sideroblasts.
Leukemia. 2011;26:542–545.
14. Patnaik MM, Lasho TL, Finke CM, et al. Spliceosome mutations involving SRSF2, SF3B1, and U2AF35 in chronic myelomonocytic leukemia:
prevalence, clinical correlates, and prognostic relevance. Am J Hematol.
2013;88(3):201–206.
15.Abu Kar S, Jankowska A, Makishima H, et al. Spliceosomal gene mutations are frequent events in the diverse mutational spectrum of chronic
myelomonocytic leukemia but largely absent in juvenile myelomonocytic
leukemia. Haematologica. 2013;98:107–113.
16.Hirabayashi S, Flotho C, Moetter J, et al. Spliceosomal gene aberrations
are rare, coexist with oncogenic mutations, and are unlikely to exert a
driver effect in childhood MDS and JMML. Blood. 2012;119(11):e96–99.
17. Wu SJ, Kuo YY, Hou HA, et al. The clinical implication of SRSF2 mutation in patients with myelodysplastic syndrome and its stability during disease evolution. Blood. 2012;120:3106–3111.
18.Lasho TL, Jimma T, Finke CM, et al. SRSF2 mutations in primary
myelofibrosis: significant clustering with IDH mutations and independent association with inferior overall and leukemia-free survival. Blood.
2012;120:4168–4171.
19. Landau DA, Carter SL, Stojanov P, et al. Evolution and impact of subclonal
mutations in chronic lymphocytic leukemia. Cell. 2013;152(4):714–726.
20.Rossi D, Bruscaggin A, Spina V, et al. Mutations of the SF3B1 splicing
factor in chronic lymphocytic leukemia: association with progression and
fludarabine-refractoriness. Blood. 2011;118:6904–6908.
21.Dreger P, Schnaiter A, Zenz T, et al. TP53, SF3B1, and NOTCH1
mutations and outcome of allotransplantation for chronic lymphocytic leukemia: six-year follow-up of the GCLLSG CLL3X trial. Blood.
2013;121(16):3284–3288.
22.Wang L, Lawrence MS, Wan L, et al. SF3B1 and other novel
cancer genes in chronic lymphocytic leukemia. N Engl J Med.
2011;365(26):2497–2506.
Vol. 105, Issue 20 | October 16, 2013
23. Quesada V, Conde L, Villamor N, et al. Exome sequencing identifies recurrent mutations of the splicing factor SF3B1 gene in chronic lymphocytic
leukemia. Nat Genet. 2012;44(1):47–54.
24.Oscier DG, Rose-Zerilli MJ, Winkelmann N, et al. The clinical significance of NOTCH1 and SF3B1 mutations in the UK LRF CLL4 trial.
Blood. 2013;121(3):468–475.
25.Quesada V, Ramsay AJ, Lopez-Otin C. Chronic lymphocytic leukemia
with SF3B1 mutation. N Engl J Med. 2012;366(26):2530.
26. Brown PJ, Ashe SL, Leich E, et al. Potentially oncogenic B-cell activationinduced smaller isoforms of FOXP1 are highly expressed in the activated B
cell-like subtype of DLBCL. Blood. 2008;111(5):2816–2824.
27.Molenaar JJ, Koster J, Zwijnenburg DA, et al. Sequencing of neuroblastoma identifies chromothripsis and defects in neuritogenesis genes. Nature.
2012;483(7391):589–593.
28.Biankin AV, Waddell N, Kassahn KS, et al. Pancreatic cancer
genomes reveal aberrations in axon guidance pathway genes. Nature.
2012;491(7424):399–405.
29.Ellis MJ, Ding L, Shen D, et al. Whole-genome analysis informs breast
cancer response to aromatase inhibition. Nature. 2012;486(7403):353–360.
30. Stephens PJ, Tarpey PS, Davies H, et al. The landscape of cancer genes and
mutational processes in breast cancer. Nature. 2012;486(7403):400–404.
31.Banerji S, Cibulskis K, Rangel-Escareno C, et al. Sequence analysis
of mutations and translocations across breast cancer subtypes. Nature.
2012;486(7403):405–409.
32.Cancer Genome Atlas Network. Comprehensive molecular portraits of
human breast tumours. Nature. 2012;490(7418):61–70.
33.Harbour JW, Roberson ED, Anbunathan H, et al. Recurrent mutations
at codon 625 of the splicing factor SF3B1 in uveal melanoma. Nat Genet.
2013;45(2):133–135.
34. Harbour JW, Onken MD, Roberson ED, et al. Frequent mutation of BAP1
in metastasizing uveal melanomas. Science. 2010;330(6009):1410–1413.
35. Krauthammer M, Kong Y, Ha BH, et al. Exome sequencing identifies recurrent somatic RAC1 mutations in melanoma. Nat Genet. 2012;44(9):1006–1014.
36. Schilling B, Bielefeld N, Sucker A, et al. Lack of SF3B1 R625 mutations in
cutaneous melanoma. Diagn Pathol. 2013;8(1):87.
37.Imielinski M, Berger AH, Hammerman PS, et al. Mapping the hallmarks of lung adenocarcinoma with massively parallel sequencing. Cell.
2012;150(6):1107–1120.
38. Schoenberg DR, Maquat LE. Regulation of cytoplasmic mRNA decay. Nat
Rev Genet. 2012;13(4):246–259.
39.An M, Henion PD. The zebrafish sf3b1b460 mutant reveals differential
requirements for the sf3b1 pre-mRNA processing gene during neural crest
development. Int J Dev Biol. 2012;56(4):223–237.
40.Isono K, Mizutani-Koseki Y, Komori T, et al. Mammalian Polycombmediated repression of Hox genes requires the essential spliceosomal protein Sf3b1. Genes Dev. 2005;19:536–541.
41.Visconte V, Rogers HJ, Singh J, et al. SF3B1 haploinsufficiency leads
to formation of ring sideroblasts in myelodysplastic syndromes. Blood.
2012;120:3173–3186.
42.Bonnal S, Vigevani L, Valcarcel J. The spliceosome as a target of novel
antitumour drugs. Nat Rev Drug Discov. 2012;11(11):847–859.
43.Kotake Y, Sagane K, Owa T, et al. Splicing factor SF3b as a target of
the antitumor natural product pladienolide. Nat Chem Biol. 2007;3(9):
570–575.
44. Kaida D, Motoyoshi H, Tashiro E, et al. Spliceostatin A targets SF3b and
inhibits both splicing and nuclear retention of pre-mRNA. Nat Chem Biol.
2007;3(9):576–583.
45. Gao Y, Koide K. Chemical perturbation of Mcl-1 pre-mRNA splicing to
induce apoptosis in cancer cells. ACS Chem Biol. 2013;8(5):895–900.
46.Yokoi A, Kotake Y, Takahashi K, et al. Biological validation that
SF3b is a target of the antitumor macrolide pladienolide. FEBS J.
2011;278(24):4870–4880.
47. Hua Y, Sahashi K, Hung G, et al. Antisense correction of SMN2 splicing
in the CNS rescues necrosis in a type III SMA mouse model. Genes Dev.
2010;24(15):1634–1644.
48. Jenny A, Hauri HP, Keller W. Characterization of cleavage and polyadenylation specificity factor and cloning of its 100-kilodalton subunit. Mol
Cell Biol. 1994;14:8183–8190.
jnci.oxfordjournals.org
49. Rosner A, Rinkevich B. The DDX3 subfamily of the DEAD box helicases:
divergent roles as unveiled by studying different organisms and in vitro
assays. Curr Med Chem. 2007;14:2517–2525.
50. Hutten S, Kehlenbach RH. CRM1-mediated nuclear export: to the pore
and beyond. Trends Cell Biol. 2007;17:193–201.
51.Ramsay AJ, Rodriguez D, Villamor N, et al. Frequent somatic mutations
in components of the RNA processing machinery in chronic lymphocytic
leukemia. Leukemia. 2013;27(7):1600-1603.
52.Raaijmakers MH, Mukherjee S, Guo S, et al. Bone progenitor dysfunction induces myelodysplasia and secondary leukaemia. Nature.
2010;464(7290):852–857.
53. Guo S, Lu J, Schlanger R, et al. MicroRNA miR-125a controls hematopoietic stem cell number. Proc Natl Acad Sci U S A. 2010;107(32):14229–14234.
54.Boocock GR, Morrison JA, Popovic M, et al. Mutations in SBDS are
associated with Shwachman-Diamond syndrome. Nat Genet. 2003;33(1):
97–101.
55. Santamaria C, Muntion S, Roson B, et al. Impaired expression of DICER,
DROSHA, SBDS and some microRNAs in mesenchymal stromal cells
from myelodysplastic syndrome patients. Haematologica. 2012;97(8):
1218–1224.
56. Kung AL, Rebel VI, Bronson RT, et al. Gene dose-dependent control of
hematopoiesis and hematologic tumor suppression by CBP. Genes Dev.
2000;14(3):272–277.
57.Rebel VI, Kung AL, Tanner EA, et al. Distinct roles for CREB-binding
protein and p300 in hematopoietic stem cell self-renewal. Proc Natl Acad
Sci U S A. 2002;99(23):14789–14794.
58.Zimmer SN, Zhou Q, Zhou T, et al. Crebbp haploinsufficiency in mice
alters the bone marrow microenvironment, leading to loss of stem cells and
excessive myelopoiesis. Blood. 2011;118(1):69–79.
59.Sobulo OM, Borrow J, Tomek R, et al. MLL is fused to CBP, a histone
acetyltransferase, in therapy-related acute myeloid leukemia with a t(11;16)
(q23;p13.3). Proc Natl Acad Sci U S A. 1997;94(16):8732–8737.
60. Lavau C, Du C, Thirman M, et al. Chromatin-related properties of CBP
fused to MLL generate a myelodysplastic-like syndrome that evolves into
myeloid leukemia. EMBO J. 2000;19(17):4655–4664.
61. Vardiman JW, Thiele J, Arber DA, et al. The 2008 revision of the World
Health Organization (WHO) classification of myeloid neoplasms and acute
leukemia: rationale and important changes. Blood. 2009;114(5):937–951.
62.Lemieux ME, Cheng Z, Zhou Q, et al. Inactivation of a single copy of
Crebbp selectively alters pre-mRNA processing in mouse hematopoietic
stem cells. PLoS One. 2011;6(8):e24153.
63. Zimmer SN, Lemieux ME, Karia BP, et al. Mice heterozygous for CREB
binding protein are hypersensitive to gamma-radiation and invariably develop myelodysplastic/myeloproliferative neoplasm. Exp Hematol.
2012;40(4):295–306 e5.
64. Denli AM, Tops BB, Plasterk RH, et al. Processing of primary microRNAs
by the Microprocessor complex. Nature. 2004;432(7014):231–235.
65.Kharas MG, Lengner CJ, Al-Shahrour F, et al. Musashi-2 regulates normal hematopoiesis and promotes aggressive myeloid leukemia. Nat Med.
2010;16(8):903–908.
Funding
This work was supported by funds from the University of Queensland Diamantina
Institute (to LMS) and the Greehey Children’s Cancer Research Institute (to VIR).
Notes
The authors of this manuscript take sole responsibility for the analysis and interpretation of the literature, writing of the manuscript, and the decision to submit
the manuscript for publication. The authors have no conflicts of interest.
Affiliations of authors: Diamantina Institute, and Faculty of Health
Sciences, School of Medicine, University of Queensland, Brisbane,
Queensland, Australia (LMS); Translational Research Institute, Brisbane,
Queensland, Australia (LMS); Greehey Children’s Cancer Research Institute,
Cancer Therapy and Research Center, and the Department of Cellular
and Structural Biology, University of Texas Health Sciences Center at San
Antonio (VIR).
JNCI | Review 1549