The AC133 epitope has been used as a marker for both normal and cancer stem cells from multiple tissue lineages. To identify transcription factors that regulate CD133 expression, we conducted parallel large-scale RNA interference screens in Caco-2 cancer cells that endogenously express CD133 and in engineered HEK293 cells that express CD133 from a heterologous promoter. The transcription factor AF4 was identified following a comparative analysis between the two screens. We then showed that AF4 is a promoter of CD133 transcription in multiple cancer cell lines. Knockdown of AF4 resulted in a dramatic reduction in CD133 transcript levels. Importantly, a subset of pediatric acute lymphoblastic leukemias (ALL) harbor a fusion oncogene results from a chromosomal translocation that juxtaposes the mixed-lineage leukemia (MLL) gene and the AF4 gene. An investigation of the functional role of CD133 in the MLL-AF4–dependent ALL cells revealed that CD133 was required for leukemia cell survival. Together, our findings show AF4-dependent regulation of CD133 expression, which is required for the growth of ALL cells. CD133 may therefore represent a therapeutic target in a subset of cancers. Cancer Res; 72(8); 1929–34. ©2012 AACR.

Cell surface recognition of AC133, an epitope on the pentaspan plasma membrane glycoprotein CD133, is one of the most commonly used stem cell and cancer stem cell (CSC) markers (1). Interestingly, AC133 marks cells that display increased chemo- and radioresistance, which are commonly used cancer therapies and may explain the observation that AC133 detection has been associated with a poor prognosis (1). Despite its ability to enrich stem cell and CSC populations, the regulation of CD133 expression remains poorly understood.

The use of AC133 has been called into question given recent findings showing that CD133 protein and transcript can be present in cells that do not display cell surface AC133 (2). We and others (3) have recently shown that AC133 relies on posttranslational modifications of CD133. Despite these findings, AC133 still requires upstream events, such as CD133 transcription, for its expression.

CD133 transcription is controlled by 5 alternative promoters (4). Moreover, CD133 promoter regions have been shown to be hypomethylated in numerous cancer cell lines and tumor samples with detectable levels of CD133 (5–7). In addition to the methylation status of the CD133 promoter, it has been shown that the Ras/extracellular signal-regulated kinase (ERK) signaling pathway is involved in activating CD133 transcription, but this was cell line–specific, suggesting that RAS/ERK activity is not the sole signaling pathway that controls CD133 transcription (8). Despite numerous studies that explore regulation of CD133 transcription, the transcription factors and mechanisms governing CD133 expression remain largely undefined.

Cell lines and culture protocols

Human embryonic kidney (HEK)293, Caco-2, HT29, and U-937 cell lines were obtained directly from American Type Culture Collection. OVCAR-8 cells were obtained from Benjamin Neel (University Health Network, Toronto, Ontario, Canada). SEM cells were obtained from the National Cell Culture Centre. All cell lines were grown in Dulbecco's Modified Eagles' Media (DMEM) supplemented with 10% FBS with antibiotics and authenticated by short tandem repeat analysis.

RNA interference knockdowns and pooled short hairpin RNA screens

The short hairpin RNA (shRNA) constructs were from the TRC lentiviral libraries available through Sigma-Aldrich Inc. (Supplementary Table S2). Lentivirus production and infection were conducted as previously described (9). Pooled lentiviral shRNA screening, shRNA-barcode probe preparation, microarray analysis, and cell sorting were conducted as previously described (10). The shRNAs were considered above background if the median probe intensity from 3 independent replicates was 256 or above.

Flow cytometry

Flow cytometric analysis was conducted as previously described (10) and was conducted to analyze Annexin-V (A13201, Molecular Probes).

Western blotting

Western blotting was conducted as previously described (10) and immunoblotting was conducted for AF4 (#A302–345A, Bethyl Laboratories Inc.).

Quantitative PCR

Quantitative PCR was carried out as previously described (10) with primers that were previously reported (11) and those that are listed in Supplementary Table S3.

Chromatin immunoprecipitation

Chromatin immunoprecipitation (ChIP) was conducted using the Imprint ChIP kit (CHP1, Sigma-Aldrich Inc.) as per manufacturer's protocol using an anti-AF4 antibody (#A302–345A, Bethyl Laboratories Inc.).

Colony-forming assay

SEM colony-forming cell assays were conducted as previously described (11) with modifications described in the Supplementary Information.

Statistics

Error bars represent SD from 3 independent replicates (n = 3). *, P ≤ 0.05; #, P ≤ 0.01. P value was calculated against the control using a 2-tailed Student t test.

Comparative analysis of large-scale RNA interference screens to identify genes involved in endogenous CD133 expression

We previously engineered a HEK293 cell line to express CD133 from a lentivirus-based cassette (HEK293/AC133; ref. 10) using the Mammalian Affinity Purification and Lentiviral Expression (MAPLE) system (12). In contrast to the HEK293/AC133 cell line, CD133 expression occurs from its native locus in the colon adenocarcinoma cell line Caco-2 (Supplementary Fig. S1). We reasoned that parallel RNA interference screens in HEK293/AC133 and Caco-2 cells would yield, through comparative analyses, transcriptional regulators that are important for native CD133 transcription. Thus, we conducted lentivirus-based pooled shRNA screens in both HEK293/AC133 (10) and Caco-2 cell lines to identify genes that regulate cell surface AC133 expression. The screening approach used a pool of 54,020 lentivirus-based shRNA constructs designed to target 11,248 genes (Fig. 1A). Changes in cell surface AC133 expression were monitored by fluorescent-activated cell sorting (FACS) and AC133low (bottom ∼2%) and AC133high cell populations were fractionated from library-infected HEK293/AC133 and Caco-2 populations (Fig. 1B, n = 3). The shRNA barcodes in the AC133low populations were quantified using a custom microarray platform (10).

Figure 1.

A comparative large-scale pooled lentivirus-based shRNA screening approach to identify genes that regulate endogenous CD133. A, Caco-2 and HEK293/AC133 cells were each infected with a 54K pooled lentiviral shRNA library at a multiplicity of infection (MOI) of 0.3 and incubated for 7 days to allow shRNA-targeted gene knockdown to occur. CMV, cytomegalovirus; VA, versatile affinity tag. B, cells were fractionated by FACS into AC133low and AC133high populations. C, genomic DNA (gDNA) of AC133low cells were extracted and used as template to PCR-amplify shRNA barcodes for deconvolution and quantification by a custom Affymetrix microarray chip (Supplementary Table S1). A median shRNA barcode probe intensity (n = 3) that was both detectable (≥256) and specific to the Caco-2 screen was considered an shRNA hit.

Figure 1.

A comparative large-scale pooled lentivirus-based shRNA screening approach to identify genes that regulate endogenous CD133. A, Caco-2 and HEK293/AC133 cells were each infected with a 54K pooled lentiviral shRNA library at a multiplicity of infection (MOI) of 0.3 and incubated for 7 days to allow shRNA-targeted gene knockdown to occur. CMV, cytomegalovirus; VA, versatile affinity tag. B, cells were fractionated by FACS into AC133low and AC133high populations. C, genomic DNA (gDNA) of AC133low cells were extracted and used as template to PCR-amplify shRNA barcodes for deconvolution and quantification by a custom Affymetrix microarray chip (Supplementary Table S1). A median shRNA barcode probe intensity (n = 3) that was both detectable (≥256) and specific to the Caco-2 screen was considered an shRNA hit.

Close modal

To identify genes specific to endogenous CD133 expression, we compared the results of the HEK293/AC133 screen and the Caco-2 screen and only considered shRNA hits specific to the Caco-2 screen (Fig. 1C, see Materials and Methods). This yielded 3,322 independent shRNAs (∼6% of the shRNA library) targeting 2,340 genes.

To discover biologic processes involved in endogenous CD133 expression, we analyzed our gene hit list using Ingenuity Canonical Pathway analysis and uncovered genes involved in transcription among the top biologic functions (Fisher's exact P = 3.69 × 10−21). Genes were selected for validation if two of its corresponding shRNAs occurred specifically in the Caco-2 screen and were ≥4-fold above background (Materials and Methods). Overall, 10 genes were validated with the original shRNAs that were considered hits from the Caco-2 screen using the following criteria (Fig. 2; Supplementary Fig. S2): first, we determined whether shRNA knockdown of the target gene resulted in a decrease in cell surface AC133 expression (Fig. 2, white bars) to confirm our primary screening data (Fig. 1B); second, we carried out quantitative PCR to measure CD133 transcript levels in Caco-2 cells (Fig. 2, black bars; Supplementary Fig. S2). Four of 10 gene hits (40%) including AF4, NME1, RBBP8, and TCERG1L caused a significant reduction in CD133 transcript levels using two independent shRNAs targeting each gene (Fig. 2). Interestingly, both NME1 and RBBP8 have been implicated in cancer as NME1 is frequently mutated in neuroblastoma (13) and RBBP8 has been suggested to be a tumor suppressor by acting in the same pathway as BRCA1 in transcriptional regulation and DNA repair (14). We elected to focus on the transcription factor AF4, a member of the AF4/LAF4/FMR2 (ALF) family as its shRNA, shAF4–2, gave the most dramatic reduction in CD133 transcript in the validated gene hits and is also not considered an essential gene across a large set of pooled shRNA screens with the same library (15).

Figure 2.

Validation of select gene hits identified as transcription factors. Caco-2 cells were infected with the selected shRNA hits and mean cell surface AC133 (white bars) and CD133 transcript expression (black bars) was monitored. Values are relative to the control shRNA targeting GFP. *, P ≤ 0.05; #, P ≤ 0.01.

Figure 2.

Validation of select gene hits identified as transcription factors. Caco-2 cells were infected with the selected shRNA hits and mean cell surface AC133 (white bars) and CD133 transcript expression (black bars) was monitored. Values are relative to the control shRNA targeting GFP. *, P ≤ 0.05; #, P ≤ 0.01.

Close modal

AF4 binds and regulates the CD133 gene in cancer cell lines

To ensure that the AF4 shRNAs were on target, we monitored AF4 levels by quantitative PCR (Fig. 3A) and by Western blotting (Fig. 3B) and observed a dramatic reduction in transcript and protein levels with two independent shRNAs targeting AF4 in the colon adenocarcinoma lines Caco-2 and HT29, as well as in the serous ovarian cancer cell line OVCAR-8. These results suggest that AF4 is involved in activation of CD133 transcription in multiple cancer cell lines. To test whether the regulation of AF4 on CD133 transcription is direct, we conducted a ChIP assay with an anti-AF4 antibody using Caco-2, HT29, and OVCAR-8 cross-linked cell lysates. We observed a significant enrichment of the CD133 gene with AF4-bound chromatin (Fig. 3C).

Figure 3.

AF4 promotes CD133 transcription. A, two independent shRNAs that target the AF4 transcript and identified as hits specific to the Caco-2 screen were used to infect cells. Quantitative PCR was used to monitor AF4 and CD133 transcripts normalized to actin. B, lysates from shRNA-treated Caco-2, HT29, and OVCAR-8 were analyzed by immunoblotting for endogenous AF4, CD133 (AC133/1), and actin. C, ChIP of endogenous AF4. Binding of AF4 to the CD133 gene was determined by quantitative PCR. *, P ≤ 0.05; #, P ≤ 0.01.

Figure 3.

AF4 promotes CD133 transcription. A, two independent shRNAs that target the AF4 transcript and identified as hits specific to the Caco-2 screen were used to infect cells. Quantitative PCR was used to monitor AF4 and CD133 transcripts normalized to actin. B, lysates from shRNA-treated Caco-2, HT29, and OVCAR-8 were analyzed by immunoblotting for endogenous AF4, CD133 (AC133/1), and actin. C, ChIP of endogenous AF4. Binding of AF4 to the CD133 gene was determined by quantitative PCR. *, P ≤ 0.05; #, P ≤ 0.01.

Close modal

Insights into AF4 mode of action at the CD133 promoter come from previous studies identifying a higher order complex containing AF4 and ENL family proteins with the positive transcription elongation factor b (PTEF-b), otherwise known as the AEP complex. AEP is recruited by transcription factors to facilitate oncogenic and physiologic mixed-lineage leukemia (MLL)-dependent transcription (16). DOT1L, the predominant histone H3 lysine 79 (H3K79) methyltransferase, is also part of AEP, and dimethylated histone H3K79 (H3K79me2) has been associated with genes that are actively transcribed. Previous studies have shown that histone H3 associated with the CD133 promoter is dimethylated at K79 (17). AF4 is not only important for transcriptional elongation of CD133 but may also promote H3K79 dimethylation through recruitment and activation of DOT1L at the CD133 locus. In support of this idea, global H3K79 methylation profiling was shown to correlate with MLL-AF4–associated gene expression in both murine and human MLL-AF4–driven leukemias (18).

Importantly, knockdown of the AEP complex subunits AF9, ENL, CDK9, CCNT1, and DOT1L also resulted in a significant decrease in CD133 transcript levels in Caco-2 cells (Supplementary Fig. S3), suggesting that AF4 functions through the AEP complex to activate CD133 transcription. The identification of TCERG1L in our screens, which has a predicted function in transcriptional elongation, also highlights the importance of transcription elongation as a major control point for CD133 expression.

MLL-AF4–regulated CD133 drives ALL cell survival

AF4 has been identified as the most frequent fusion with the MLL gene due to a chromosomal translocation in pediatric acute lymphoblastic leukemias (ALL) and is associated with a particularly poor prognosis (11). Interestingly, MLL-AF4 has been previously shown to bind to the CD133 gene (17), and its knockdown causes downregulation of CD133 transcript and cell surface CD133 expression (11).

The MLL gene located on chromosome 11 band q23 is frequently involved in chromosomal translocations resulting in fusions with not only AF4 but also other members of the AEP complex (e.g., AF9 or ENL). Given our findings that the AEP complex is required to activate CD133 transcription (Supplementary Fig. S3), we speculate that MLL fusions, in addition to AF4, may also act to promote CD133 transcription in ALLs. Indeed, a previous study reported that all samples of 8 patients with pro-B-ALL containing an MLL gene translocation expressed cell surface AC133, whereas only 2 of the 9 patients with pro-B-ALL samples lacking an MLL gene translocation expressed cell surface AC133 (19). Consistent with this finding, AF4 knockdown in the human leukemic monocyte lymphoma cell line U-937, which lacks an MLL fusion oncogene, did not result in a reduction in CD133 transcript levels (Supplementary Fig. S4). Importantly, knockdown of MLL-AF4 resulted in a dramatic reduction of CD133 transcript (Fig. 4A) and cell surface AC133 detection in SEM cells (Fig. 4D). Moreover, knockdown of AEP complex subunits also resulted in a significant downregulation of CD133 transcript in SEM cells (Supplementary Fig. S3a).

Figure 4.

MLL-AF4–driven CD133 is required for leukemia progression. A, knockdown of the MLL-AF4 transcript in SEM by shAF4–1 was determined by quantitative PCR and normalized to actin. B, actin-normalized HOXA7 and MEIS1 levels were monitored in the knockdown of CD133 by quantitative PCR in SEM cells. C, a methylcellulose assay of shRNA-treated SEM cells to assess colony-forming units. D, cell surface AC133 (CD133) expression and Annexin-V were monitored by FACS of shRNA-treated SEM cells. APC, allophycocyanin. *, P ≤ 0.05; #, P ≤ 0.01.

Figure 4.

MLL-AF4–driven CD133 is required for leukemia progression. A, knockdown of the MLL-AF4 transcript in SEM by shAF4–1 was determined by quantitative PCR and normalized to actin. B, actin-normalized HOXA7 and MEIS1 levels were monitored in the knockdown of CD133 by quantitative PCR in SEM cells. C, a methylcellulose assay of shRNA-treated SEM cells to assess colony-forming units. D, cell surface AC133 (CD133) expression and Annexin-V were monitored by FACS of shRNA-treated SEM cells. APC, allophycocyanin. *, P ≤ 0.05; #, P ≤ 0.01.

Close modal

Similar to a previous study, we observed a significant decrease in HOXA7 and MEIS1 when suppressing MLL-AF4 in SEM cells (Fig. 4B; ref. 11). We also observed a decrease in HOXA7 and MEIS1 following CD133 knockdown (Fig. 4B). As leukemia is often characterized by the accumulation of hematopoietic cells that are unable to differentiate into functional progeny, our findings suggest that a critical function of MLL-AF4 is to upregulate CD133 to suppress cell differentiation. Indeed, this is consistent with the observation that detection of AC133 at the cell surface of ALL cells has been shown to mark the primitive leukemia-initiating cells responsible for leukemia progression (20).

To support the hypothesis that CD133 functions to maintain SEM cells in a primitive state, we investigated the impact of CD133 on the capacity of SEM cells to form colonies in methylcellulose, an in vitro measurement of leukemia potential. CD133 was knocked down with two independent shRNAs in SEM cells and these cells were cultured in methylcellulose to observe colony formation. A significant decrease in the number of colonies was observed in MLL-AF4 and CD133-depleted samples compared with control cells expressing a nontargeting shRNA (Fig. 4C).

To investigate whether the reduction of SEM colony formation due to CD133 knockdown resulted from apoptosis, cell surface Annexin-V levels were examined in SEM cells following knockdown of CD133 using 2 independent shRNAs. CD133 or MLL-AF4 depletion resulted in cell surface Annexin-V staining (Fig. 4D), indicating that SEM cells require CD133 for cell survival. Interestingly, a previous study showed that targeting CD133 in the metastatic melanoma cell line FEMX-I also resulted in apoptosis (21), showing that the requirement for CD133 for cell survival is not restricted to hematopoietic malignancies. These observations imply that CD133 could serve as a useful therapeutic target in certain cancers.

No potential conflicts of interests were disclosed.

The authors thank D. Kasimer and P.-A. Penttila for technical assistance.

A.B. Mak was supported by an NSERC CGS-D. This work was supported by funds to J. Moffat from SCN, CIHR, and OICR. J. Moffat holds a CIHR New Investigator award and is a CIFAR scholar.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Ferrandina
G
,
Petrillo
M
,
Bonanno
G
,
Scambia
G
. 
Targeting CD133 antigen in cancer
.
Expert Opin Ther Targets
2009
;
13
:
823
37
.
2.
Bidlingmaier
S
,
Zhu
X
,
Liu
B
. 
The utility and limitations of glycosylated human CD133 epitopes in defining cancer stem cells
.
J Mol Med
2008
;
86
:
1025
32
.
3.
Kemper
K
,
Sprick
MR
,
de
BM
,
Scopelliti
A
,
Vermeulen
L
,
Hoek
M
, et al
The AC133 epitope, but not the CD133 protein, is lost upon cancer stem cell differentiation
.
Cancer Res
2010
;
70
:
719
29
.
4.
Shmelkov
SV
,
Jun
L
,
St
CR
,
McGarrigle
D
,
Derderian
CA
,
Usenko
JK
, et al
Alternative promoters regulate transcription of the gene that encodes stem cell surface protein AC133
.
Blood
2004
;
103
:
2055
61
.
5.
Baba
T
,
Convery
PA
,
Matsumura
N
,
Whitaker
RS
,
Kondoh
E
,
Perry
T
, et al
Epigenetic regulation of CD133 and tumorigenicity of CD133+ ovarian cancer cells
.
Oncogene
2009
;
28
:
209
18
.
6.
Tabu
K
,
Sasai
K
,
Kimura
T
,
Wang
L
,
Aoyanagi
E
,
Kohsaka
S
, et al
Promoter hypomethylation regulates CD133 expression in human gliomas
.
Cell Res
2008
;
18
:
1037
46
.
7.
Yi
JM
,
Tsai
HC
,
Glockner
SC
,
Lin
S
,
Ohm
JE
,
Easwaran
H
, et al
Abnormal DNA methylation of CD133 in colorectal and glioblastoma tumors
.
Cancer Res
2008
;
68
:
8094
103
.
8.
Tabu
K
,
Kimura
T
,
Sasai
K
,
Wang
L
,
Bizen
N
,
Nishihara
H
, et al
Analysis of an alternative human CD133 promoter reveals the implication of Ras/ERK pathway in tumor stem-like hallmarks
.
Mol Cancer
2010
;
9
:
39
.
9.
Moffat
J
,
Grueneberg
DA
,
Yang
X
,
Kim
SY
,
Kloepfer
AM
,
Hinkle
G
, et al
A lentiviral RNAi library for human and mouse genes applied to an arrayed viral high-content screen
.
Cell
2006
;
124
:
1283
98
.
10.
Mak
AB
,
Blakely
KM
,
Williams
RA
,
Penttila
PA
,
Shukalyuk
AI
,
Osman
KT
, et al
CD133 N-glycosylation processing contributes to cell-surface recognition of the primitive cell marker AC133
.
J Biol Chem
2011
;
286
:
41046
56
.
11.
Thomas
M
,
Gessner
A
,
Vornlocher
HP
,
Hadwiger
P
,
Greil
J
,
Heidenreich
O
. 
Targeting MLL-AF4 with short interfering RNAs inhibits clonogenicity and engraftment of t(4;11)-positive human leukemic cells
.
Blood
2005
;
106
:
3559
66
.
12.
Mak
AB
,
Ni
Z
,
Hewel
JA
,
Chen
GI
,
Zhong
G
,
Karamboulas
K
, et al
A lentiviral functional proteomics approach identifies chromatin remodeling complexes important for the induction of pluripotency
.
Mol Cell Proteomics
2010
;
9
:
811
23
.
13.
De
PK
,
Vandesompele
J
,
Heimann
P
,
Yigit
N
,
Beckman
S
,
Schramm
A
, et al
Human fetal neuroblast and neuroblastoma transcriptome analysis confirms neuroblast origin and highlights neuroblastoma candidate genes
.
Genome Biol
2006
;
7
:
R84
.
14.
You
Z
,
Bailis
JM
. 
DNA damage and decisions: CtIP coordinates DNA repair and cell cycle checkpoints
.
Trends Cell Biol
2010
;
20
:
402
9
.
15.
Marcotte
R
,
Brown
KR
,
Suarez
F
,
Sayad
A
,
Karamboulas
K
,
Krzyzanowski
PM
, et al
Essential gene profiles in breast, pancreas and ovarian cancer cells
.
Cancer Discovery
2012
;
2
:
95
6
.
16.
Yokoyama
A
,
Lin
M
,
Naresh
A
,
Kitabayashi
I
,
Cleary
ML
. 
A higher-order complex containing AF4 and ENL family proteins with P-TEFb facilitates oncogenic and physiologic MLL-dependent transcription
.
Cancer Cell
2010
;
17
:
198
212
.
17.
Guenther
MG
,
Lawton
LN
,
Rozovskaia
T
,
Frampton
GM
,
Levine
SS
,
Volkert
TL
, et al
Aberrant chromatin at genes encoding stem cell regulators in human mixed-lineage leukemia
.
Genes Dev
2008
;
22
:
3403
8
.
18.
Krivtsov
AV
,
Feng
Z
,
Lemieux
ME
,
Faber
J
,
Vempati
S
,
Sinha
AU
, et al
H3K79 methylation profiles define murine and human MLL-AF4 leukemias
.
Cancer Cell
2008
;
14
:
355
68
.
19.
Wuchter
C
,
Ratei
R
,
Spahn
G
,
Schoch
C
,
Harbott
J
,
Schnittger
S
, et al
Impact of CD133 (AC133) and CD90 expression analysis for acute leukemia immunophenotyping
.
Haematologica
2001
;
86
:
154
61
.
20.
Cox
CV
,
Diamanti
P
,
Evely
RS
,
Kearns
PR
,
Blair
A
. 
Expression of CD133 on leukemia-initiating cells in childhood ALL
.
Blood
2009
;
113
:
3287
96
.
21.
Rappa
G
,
Fodstad
O
,
Lorico
A
. 
The stem cell-associated antigen CD133 (Prominin-1) is a molecular therapeutic target for metastatic melanoma
.
Stem Cells
2008
;
26
:
3008
17
.