Search tips
Search criteria 


Logo of plosonePLoS OneView this ArticleSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)
PLoS One. 2013; 8(8): e70812.
Published online 2013 August 14. doi:  10.1371/journal.pone.0070812
PMCID: PMC3743845

The Prognostic Significance of IRF8 Transcripts in Adult Patients with Acute Myeloid Leukemia

Zhuang Zuo, Editor


Interferon regulatory factor 8 (IRF8) is a transcription factor that plays a critical role in normal hematopoiesis, such that disruption of IRF8 activity promotes leukemogenesis. We and others have identified aberrant expression of IRF8 transcripts, including novel splice variants, in acute myeloid leukemia (AML), but studies have not investigated the prognostic significance of these transcripts. Therefore, we developed and optimized quantitative expression assays for both, the wild type, or the reference sequence (WT-IRF8) and novel splice variants (SV-IRF8). These assays were used to quantify IRF8 transcript levels in 194 adult patients with AML, and multivariate analyses investigated the prognostic significance of these expression levels. After adjusting for known prognostic factors, expression levels of WT- or SV-IRF8 transcripts were not significantly associated with complete responses or overall survival. However, increased expression of WT-IRF8 was associated with decreased relapse-free survival (RFS) in both univariate (P = 0.010) and multivariate (P = 0.019) analyses. Similarly, increased expression of SV-IRF8 was associated with a decreased RFS (univariate, P = 0.026 and multivariate, P = 0.021). These studies show for the first time that WT-IRF8 and SV-IRF8 are independent adverse prognostic factors for patients with AML. Additional studies are planned to examine the prognostic significance of IRF8 transcripts in other populations of AML patients.


Interferon regulatory factor 8 (IRF8) is a tightly regulated transcription factor [1], [2]. Modest changes in IRF8 activity help to determine hematopoietic lineage, and disruptions in this activity detrimentally impacts normal hematopoiesis [1], [2]. For example, loss of IRF8 function disproportionately promotes the development of granulocytes and blocks normal monocytic differentiation. Conversely, constitutive expression of IRF8 promotes the development of monocytic/dendritic-like cells and blocks normal granulocytic differentiation [3][5]. Over time, IRF8-related dysregulation may promote malignant transformation, and in fact, inactivation of IRF8 in mice models causes a myeloproliferative syndrome that progresses to an AML-like disease [6][8].

We and others have shown that aging hematopoietic stem cells and primary leukemic blasts frequently display aberrant expression of the wild type reference sequence for IRF8 (WT-IRF8, NM_144701) [9][15]. In addition, we have recently identified novel splice variants for IRF8 (SV-IRF8) in malignant cell lines and leukemic blasts from AML patients, which are subsequently described in this manuscript. Although the functional significance of the splice variants remains unknown, these SV-IRF8 transcripts appear to be the result of aberrant hypermethylation of the normal IRF8 promoter, and as such, may not be subject to the normal regulatory factors that control WT-IRF8 expression (unpublished data).

The available data indicate that normal IRF8 activity is essential for a healthy hematopoietic system and that disruption of this activity may promote leukemogenesis. Despite this data, studies have not examined the prognostic significance of IRF8 transcripts in AML or other hematopoietic malignancies. Therefore, we investigated the prognostic significance of IRF8 transcripts in a large cohort of adult patients with AML who were treated on Southwest Oncology Group (SWOG) protocols. These studies showed for the first time that IRF8 transcripts were independent adverse prognostic biomarkers for adult patients with AML.

Materials and Methods

Ethics Statement

All participants provided written informed consent in compliance with the Declaration of Helsinki to permit the use of their samples for research. Documentation of consent was provided to and maintained by SWOG, and only samples from consented subjects were included in the studies. All IRF8 studies were conducted with approval of the Fred Hutchinson Cancer Research Center's Institution Review Board, which oversees the ethical conduct of research at this center.

Patient material

AML samples used for development of WT-IRF8 and SV-IRF8 assays were obtained from the Fred Hutchinson Cancer Research Center and the University of Washington Leukemia Repository. Peripheral blood from normal donors was obtained under FHCRC IRB-approved protocol. Cell lines were purchased from ATCC (Manassas, VA). RNA from diagnostic bone marrow (BM, N = 155) or peripheral blood (PB, N = 39) samples were obtained from 194 previously untreated AML patients who received cytarabine (Ara-C) and daunorubicin (DNR)-based induction as part of SWOG-9031, SWOG-9126, SWOG-9333 or SWOG-9500 studies [16]-[19]. Cytogenetic studies, centrally reviewed by the SWOG Cytogenetics Committee, were available for most patients. FLT3, NPM1, and DNMT3A molecular studies for the samples have previously been described, and results were obtained from the SWOG database [20][22].

Identification of Novel Splice Variants

The WT-IRF8 sequence was obtained from the UCSC Human Genome Browser (NM_144701; g=refGene&i=NM_144701). Qualitative RT/PCR assays were developed to examine the entire coding region of the WT-IRF8 transcript (Table S1). Primer IRF8.Ex1.F annealed to the 5′ untranslated region (UTR) within exon 1, while primer IRF8.Ex2.F) annealed to exon 2, which contains the start codon. A universal reverse primer (IRF8.Ex9.R) annealed to the 3′ UTR in exon 9 distal to the stop codon. The sequencing primers were scattered approximately every 300–400 Bps along the coding region (Table S1). cDNA was generated using poly T primers, standard reagents, and conditions as previously described. The entire WT-IRF8 transcript was amplified using High Fidelity Platinum Taq polymerase (Invitrogen, Carlsbad, CA, USA) and conditions provided in Table S1. GeneRacerTM kit (Invitrogen) was utilized for RNA ligase-mediated rapid amplification of 5′ end of IRF8 transcripts using the GeneRacer™ 5′ Primer (forward) and either IRF8.Ex3.R or IRF8.Ex9.GR.R primers (Table S1). Amplified products were then sequenced using the amplification primer set and conditions provided in Table S1.

Development of Q-RT/PCR Assays for WT-IRF8 and SV-IRF8

Based on the consensus sequence of the three SV-IRF8 transcripts (File S1, IRF8 Sequences), a forward primer within the cryptic 1st exon was developed to amplify all three SVs. Two probes for quantitative real-time PCR (Q-PCR) assays were developed, P1 and P2. P1 annealed to a unique sequence spanning exons 1 and 2 of the WT-IRF8 transcript. When coupled with primers F1 and R1, P1 was used to quantify WT-IRF8 transcripts by Q-PCR. P2 annealed across the junction of exons 2 and 3, and when coupled with primers F2 and R2, P2 was used to quantify the total expression of all three SV-IRF8 transcripts. Primer and probe sequences are provided in Table S2 and are illustrated in Fig 1.

Figure 1
Q-RT/PCR assay design for WT- and SV-IRF8 transcripts.

WT-IRF8, SV-IRF8, CD34 and GUSB (e.g., housekeeping control) assays were performed in duplicate using cDNA generated from RNA as previously described [23]. Primers, probes, and reagents were mixed and subjected to 45 cycles of amplification using standard thermal cycling conditions on Applied Biosystems 7900HT Fast Real-Time PCR system [23]. All experiments contained appropriate negative and positive controls, including a calibrator sample, i.e., a pool of RNA from peripheral blood mononuclear cells of 10 normal donors. The fold differences were computed using 2−ΔΔCT method, employing GUSB expression to correct for RNA integrity and the PB expression as the control sample for computation of fold change as previously described [24].

Development of Fragment Analysis Assay for Individual SV-IRF8 Transcripts

A fragment analysis assay was developed utilizing the same primers that were used for Q-RT/PCR SV-IRF8 assay (i.e., F2 and R2). The forward primer (F2), which annealed to a conserved sequence within the cryptic 1st exon, was fluorescently labeled with FAM. The fragment analysis assay was optimized and performed using cDNA generated from RNA. Primers (FAM-F2 and R2) and standard reagents were mixed and subjected to 35 cycles of amplification (Table S1, Fragment Analysis). Analyses were performed using the GeneMapper software as previously described [21].

Statistical methods

Q-RT/PCR assays for WT-IRF8, SV-IRF8 transcripts, CD34 and GUSB were performed in duplicate. Geometric means of the duplicate fold-change values were used for all statistical analyses (File S2, Additional Statistical Methods). Data regarding patient characteristics and treatment outcomes, including complete response (CR), resistant disease (RD), overall survival (OS), relapse-free survival (RFS), and CR duration were collected and evaluated according to standard SWOG procedures [16][19]. OS was measured from the date of the patient's entry into the clinical trial until death from any cause, with observation censored at the date of last contact for patients not known to have died. RFS and CR duration were measured from the date of CR until AML relapse or death from any cause with censoring at the date of last contact for patients with no report of relapse (RFS) or until relapse with censoring at the last contact or death (CR duration). Co-factors for statistical analyses of outcome included FLT3 internal tandem duplications (FLT3-ITD), NPM1 mutation, DNMT3A mutation, age, sex, race, white blood cell (WBC) count, BM and PB blast percentages at study entry, FAB classification, AML onset (i.e., de novo vs. secondary), and cytogenetic risk group. Associations between continuous variables were measured by Spearman's rank order correlation coefficient (ρ). Comparisons between groups of patients were based on Wilcoxon or Kruskal-Wallis rank sum tests for continuous variables, and on logistic and proportional hazards regression analyses for dichotomous and time-to-event data, respectively. Statistical significance was represented by two-sided p-values (P).


Identification of Novel Splice Variants

Potential loss of exon 1 was suggested by attempts to amplify the entire sequence of IRF8 in cell lines. Many of evaluated cell lines showed amplification using Ex2.F/Ex9.R primer set (Fig 2, second panel) but not Ex1.F/Ex9.R (Fig 2, first panel). Using the GeneRacer Kit, three novel splice variants were identified and sequenced. In each case, a cryptic 1st exon originating from within the normal intron 1 region was detected (File S1, IRF8 Sequences), while exon 1 sequence was missing. A forward primer annealing to all the splice variants was developed and used to confirm that the alternative splice variants retained the entire coding region of IRF8 (i.e., exons 2–9, Fig 2, panel 3). For two of the splice variants, a potential start codon was identified within the cryptic exon 1 and prior to the normal start codon (Figs 1 and and2,2, File S1, IRF8 Sequences). It is unknown if these two splice variants begin translation at the alternative start codon, but preliminary studies demonstrate that the cell lines that primarily transcribe SV-IRF8 transcripts express an IRF8 protein (File S3, Protein Expression).

Figure 2
Identification of Novel IRF8 Splice Variants.

Characterization of patient population

The characteristics of 194 AML patients with available material were compared to 530 subjects from the same SWOG trials who were excluded due to an induction regimen other than Ara-C (Days 1–7) and DNR (Days 1–3), or absence of material for investigation. The 194 included patients had higher WBC counts and blast percentages. In addition, the included patients were younger, more likely to have normal karyotype, and less likely to have deletions of chromosomes 5q and/or 7q (Table 1). After adjusting for these factors, the 194 included patients did not differ significantly in CR or RD rates, RFS, CR duration, or OS from other trial participants.

Table 1
Characteristics of patients with IRF8 data and excluded patients.

Associations with other clinical and molecular factors

WT-IRF8 was expressed at low levels in majority of AML patients (range 0.00728–8.317, Fig 3A), but 12% (24/194) displayed a marked increased WT-IRF8 expression (>2-fold). SV-IRF8 was also expressed at low levels in most AML patients (range 0.002–115.031, Fig 3B), but, a subgroup of AML patients (26/192, 14%) also displayed >2-fold increase in SV-IRF8 expression. WT-IRF8 and SV-IRF8 expression were significantly correlated (ρ = 0.52, P<0.0001); however, many samples with increased SV-IRF8 expression displayed very low levels of WT-IRF8 and vice versa (Fig 3C). To determine if mutations within the coding sequence of WT-IRF8 may be present in samples with increased expression of IRF8, ten samples with high expression of both SV- and WT-IRF8 were sequenced using multiple primers (Table S1). None of these samples harbored mutations within the coding exons of the gene. Increased expression of WT-IRF8 was significantly associated with FAB M5 phenotype (acute monoblastic or monocytic leukemia, P<0.0001). In addition, increased expression of WT-IRF8 was modestly associated with higher CD34 expression (ρ = 0.20, P = 0.0046), lower PB blast percentage (ρ = −0.18, P = 0.011), and an absence of FLT3-ITD (P = 0.027, Tables 2 and and3).3). WT-IRF8 expression was not significantly associated with other potential prognostic factors, including age, other molecular biomarkers (NPM1 or DNMT3A mutations) or cytogenetic risk group (Tables 2 and and33).

Figure 3
Expression of IRF8 transcripts in samples from healthy donors and AML patients.
Table 2
Rank order correlation of WT-IRF8 and SV-IRF8 expression (fold-change, treated as a continuous variable) with patient characteristics, based on 194 adult patients with previously untreated AML.
Table 3
Association of WT-IRF8 expression (fold-change) with patient characteristics, based on 194 adult patients with previously untreated AML.

Similar to WT-IRF8, increased expression of SV-IRF8 was associated with higher levels of CD34 expression and lower PB blast percentage (ρ = 0.36, P<0.0001 and ρ = −0.18, P = 0.013, respectively). Increased expression of SV-IRF8 was not significantly associated with M5 FAB phenotype (P = 0.31); however, increased expression of SV-IRF8 was correlated with lower BM blast percentage (ρ = −0.18, P = 0.018) and an absence of NPM1 mutations (P = 0.039, Tables 2 and and4),4), with the later finding being particularly true for samples with >2-fold SV-IRF8 expression (P  = 0.0047). SV-IRF8 was also not significantly associated with age, other molecular biomarkers (FLT3-ITD or DNMT3A mutations) or cytogenetic risk groups (Tables 2 and and44).

Table 4
Association of SV-IRF8 expression (fold-change) with patient characteristics, based on 192 adult patients with previously untreated AML.

WT-IRF8 and SV-IRF8 as Adverse Prognostic Biomarkers

As a continuous variable, WT-IRF8 expression was not significantly associated with CR, RD, or OS; however, increasing WT-IRF8 expression was associated with a reduction in RFS (P = 0.010 in univariate analysis, P = 0.019 in multivariate analysis). Examining WT-IRF8 expression as a dichotomous variable, an increased WT-IRF8 expression (>2-fold) was associated with a lower CR rate (25%, 6/24, vs. 52%, 88/170, P = 0.012) and a reduced OS as exemplified by an increased hazard ratio (HR) of 1.63 (P = 0.042), but these adverse associations did not retain their significance after adjusting for other prognostic factors (P = 0.33 and P = 0.63, respectively). However, >2-fold expression of WT-IRF8 was significantly associated with shorter RFS in both univariate (P = 0.0099) and multivariate analyses (P = 0.011). Notably, all patients with >2-fold WT-IRF8 expression (N = 6) relapsed within 8 months after achieving CR (Fig 4A), with adjusted HR of 3.13 (95% CI 1.30–7.49). These 6 patients were not restricted to a single FAB class (i.e., 1 patient was classified as having FAB M1, another – FAB M2, two patients – FAB M4, and two more – FAB M5). These findings are summarized in Table 5. Since 58 of the 71 RFS events were relapses, the effect of WT-IRF8 on RFS was largely due to its impact on risk of relapse: >2-fold expression was associated with shorter CR duration in univariate (P = 0.0099) and multivariate (P = 0.011) analyses. Within the group of patients with >2-fold WT-IRF8 expression, patients who obtained a CR and relapsed had WT-IRF8 levels similar to those who never obtained a CR (Fig 5).

Figure 4
Kaplan-Meier estimates of Relapse-Free Survival for patients who achieved CR.
Figure 5
WT- and SV-IRF8 expression among patients with >2-fold expression, by outcome.
Table 5
Treatment outcomes by WT-IRF8 expression.

Increasing SV-IRF8 expression as a continuous variable was associated with lower CR rate (P = 0.0023), higher RD (P = 0.028) and lower OS (P = 0.051), but these associations did not retain their significance after adjusting for other prognostic factors. Similar to WT-IRF8, increasing SV-IRF8 expression was associated with reduced RFS (P = 0.026), which remained significant after adjusting for other prognostic factors (P = 0.021). Furthermore, as a dichotomous variable, >2-fold SV-IRF8 expression was also significantly associated with reduced RFS in both unadjusted (P = 0.029) and adjusted (P = 0.017) analyses (Fig 4B). Patients with >2-fold SV-IRF8 had an adjusted HR for RFS of 2.89 (95% CI 1.21–6.89). These findings are summarized in Table 6. As with WT-IRF8, the effect of SV-IRF8 expression was primarily on risk of relapse: >2-fold expression was associated with shorter CR duration in univariate (P = 0.043) and multivariate (P = 0.027) analyses. Similarly, patients with >2-fold SV-IRF8 expression who obtained a CR and relapsed did not display a markedly different median expression levels than patients who never obtained a CR (Fig 5).

Table 6
Treatment outcomes by SV-IRF8 expression.

Multivariate models were used to test whether the apparent effect of IRF8 on RFS might be due to the association of IRF8 with FAB. These analyses were limited to patients with FAB M1, M2, M4 or M5 (combined as M4 or M5 vs. M1 or M2 to enhance statistical power); the two patients with “Other” FAB due to “M4/M5” or “M4 or M5B” were included in the “M4 or M5” group. Although patients with >2-fold WT-IRF8 expression accounted for a slightly higher proportion of the M4 or M5 group (4/40 = 10%) compared to M1 or M2 (2/50 = 4%), this difference was small. This trend was also observed in the group of patients with >2-fold SV-IRF8 expression: 3/38 = 7.9% for M4 or M5 vs. 3/50 = 6% for M1 or M2. In multivariate proportional hazards regression analyses incorporating both FAB and IRF8 expression, IRF8 retained its significant effects on RFS, while FAB (M4 or M5 vs. M1 or M2) was not significant (Table S3).

Correlation between individual SV-IRF8 transcripts and clinical outcomes

Fragment analyses data were available for 165 of the 194 patients. The most abundant dominant transcript variant was SV1-IRF8 (125/165, 76%), followed by SV2-IRF8 (28/125, 17%) and SV3-IRF8 (12/165, 12%). SV1-IRF8, as a continuous variable, was associated with higher expression of WT-IRF8 (P = 0.0001), while there were no significant associations between individual splice variants and SV-IRF8 expression (Table S4). CD34 expression was somewhat higher in patients with dominant SV2-IRF8 (P = 0.0091, Table S4).

Variations in clinical outcomes for the three dominant SV transcripts were not statistically significant (CR = 0.38; RD P = 0.25; OS P = 0.077; RFS P = 0.32, Table S5). Additional analyses were performed to test whether the above-described prognostic effects for WT-IRF8 or SV-IRF8 varied significantly between the SV1- and SV2–dominant patients; the SV3 group was excluded due to its small size. In multivariate analyses, adjusting for other significant prognostic factors, there were no significant interactions between the dominant SV transcript and SV-IRF8 expression for the outcomes of CR, RD, RFS, or OS (Table S6). Likewise, there were no significant interactions between the dominant SV transcript and WT-IRF8 expression for the outcomes of CR, RD, or OS. There was a modestly significant interaction between the dominant SV transcript and the effect of WT-IRF8 on RFS, with the impact of WT-IRF8 expression on RFS being greater for SV1 compared to SV2 (P = 0.043 for continuous expression and P = 0.036 for >2–fold expression; Fig 6). However, due to the small number of SV2-dominant patients in this analysis (N = 16), this result must be interpreted with caution.

Figure 6
Kaplan-Meier estimates of Relapse-free survival by WT-IRF8 expression and dominant IRF8 splice variant (SV1 or SV2).


Although current prognostic biomarkers (e.g., FLT3-ITD, NPM1 mutations, etc.) have been helpful in risk-stratifying patients with AML, these molecular biomarkers remain somewhat unsatisfactory for accurately predicting clinical outcomes for a subset of patients with AML. Given the molecular heterogeneity and complexity of AML, it is unlikely that a single or very restricted number of prognostic biomarkers will ever display the predictive power necessary to accurately predict clinical outcomes for all AML patients. Therefore, there is a need to identify additional molecular biomarkers that can be combined with other known prognostic factors to more precisely predict clinical outcomes for patients with AML. Previous studies have not examined the prognostic significance of IRF8 transcripts [10][14]. Our results indicate that WT-IRF8 and SV-IRF8 transcripts are adverse prognostic biomarkers for predicting RFS in adult patients with AML patients. Moreover, the prognostic significance of these biomarkers was independent of other known prognostic factors. For example, after adjusting for potential associated prognostic factors (e.g., NPM1, FAB, etc.), increased expression of WT-IRF8 and SV-IRF8 transcripts remained significantly associated with an inferior RFS. Given the independent nature of these biomarkers, IRF8 transcripts, therefore, may be combined with other prognostic factors (i.e., molecular and clinical) to develop a more comprehensive risk-stratification tool for patients with AML.

Since loss of IRF8 activity has been associated with dysfunctional hematopoiesis and leukemogenesis, it was somewhat unexpected that increased WT-IRF8 expression was associated with an overall worse prognosis. However, just as a loss of IRF8 expression disrupts hematopoiesis [25], constitutive or over-expression of IRF8 may also have a detrimental effect on normal hematopoiesis. Certainly, forced over-expression of IRF8 blocks neutrophil differentiation and promotes the development of macrophages, monocytes and dendritic cells [5], [26], [27]. Although the current correlative studies were not designed to precisely characterize the functional changes associated with increased IRF8 expression, increased WT-IRF8 expression was significantly associated with FAB M5 AML (i.e., monocytic leukemia), which is consistent with previous ex vivo studies examining the functional impact of over-expression of WT-IRF8 [5], [26], [27]. Furthermore, if increased IRF8 expression causes the leukemic blasts to harbor aberrant dendritic-like qualities, these functional changes may explain some of adverse prognostic impact for increased WT-IRF8, given that classic dendritic cell leukemia/neoplasms have a particularly unfavorable prognosis [28].

With respect to SV-IRF8, it remains uncertain if SV-IRF8 transcripts are translated into proteins or may be surrogate biomarkers for some other unfavorable biological state. If the splice variants are translated into structurally different proteins than WT-IRF8, these structural differences may lead to functional changes, which could explain the negative clinical impact of the aberrant splice variants. This may be especially true if the abnormal protein(s) abrogate or disrupt the function of the WT protein, resulting in a dominant-negative phenotype. Previous studies have shown that expression of WT-IRF8 transcript is normally regulated by the methylation status of its promoter, such that hypermethylation of the promoter typically down regulates the expression of the WT-IRF8 transcript [29]. However, we have found that aberrant hypermethylation of the normal IRF8 promoter may also promote the expression of the SV-IRF8 transcripts (unpublished data). Hence, expression of the IRF8 splice variants may be a surrogate biomarker for global hypermethylation of the genome and the potential adverse consequences of this hypermethylation. Likewise, hypermethylation of normal IRF8 promoter may interfere with the normal regulatory processes controlling IRF8 transcription. In this case, even if SV-IRF8 transcripts code for the wild-type protein, the expression of the IRF8 protein may be at inappropriate times or levels, and as such, may lead to cellular dysfunction.

In conclusion, patients with increased expression of WT-IRF8 or SV-IRF8 transcripts have significantly shorter durations of remission and RFS. Studies are underway to examine the functional impact of increased WT-IRF8 expression on chemotherapy sensitivity and the coding potential of the SV-IRF8 transcripts. Furthermore, larger correlative studies examining more diverse populations of patients with AML and other hematopoietic malignancies are planned. These future correlative investigations will examine IRF8 transcripts and other molecular biomarkers across broad range of different therapies and ages, including pediatric subjects.

Supporting Information

Table S1

Qualitative PCR primers.


Table S2

Quantitative RT/PCR primers and Commercial Assays.


Table S3

The Effect of IRF8 Expression on RFS and associations with FAB.


Table S4

Association of IRF8 and CD34 expression (fold-change) with dominant IRF8 splice variant, based on 165 adult patients with previously untreated AML.


Table S5

Association of treatment outcomes with dominant IRF8 splice variant, based on 165 adult patients with previously untreated AML.


Table S6

Summary of multivariate analyses testing whether prognostic effects of WT- or SV-IRF8 differ between SV1- and SV2-IRF8 transcripts.


File S1

IRF8 Sequences.


File S2

Additional Statistical Methods.


File S3

IRF8 Protein.



The authors would like to acknowledge and thank the Fred Hutchinson Cancer Research Center and University of Washington Leukemia Repository for providing AML samples, which were utilized in the development of WT-IRF8 and SV-IRF8 assays.

Funding Statement

This study was supported in part by National Institute of Health grant numbers CA114563 (DLS), CA160872 (DLS), CA139256 (SM), and by the following PHS Cooperative Agreement grant numbers awarded by the National Cancer Institute, DHHS: CA32102 and CA38926. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.


1. Schmidt M, Bies J, Tamura T, Ozato K, Wolff L (2004) The interferon regulatory factor ICSBP/IRF-8 in combination with PU.1 up-regulates expression of tumor suppressor p15(Ink4b) in murine myeloid cells. Blood 103: 4142–4149 [PubMed]
2. Zhu C, Saberwal G, Lu Y, Platanias LC, Eklund EA (2004) The interferon consensus sequence-binding protein activates transcription of the gene encoding neurofibromin 1. J Biol Chem 279: 50874–50885 [PubMed]
3. Tsujimura H, Nagamura-Inoue T, Tamura T, Ozato K (2002) IFN consensus sequence binding protein/IFN regulatory factor-8 guides bone marrow progenitor cells toward the macrophage lineage. J Immunol 169: 1261–1269 [PubMed]
4. Tamura T, Thotakura P, Tanaka TS, Ko MS, Ozato K (2005) Identification of target genes and a unique cis element regulated by IRF-8 in developing macrophages. Blood 106: 1938–1947 [PubMed]
5. Becker AM, Michael DG, Satpathy AT, Sciammas R, Singh H, et al. (2012) IRF-8 extinguishes neutrophil production and promotes dendritic cell lineage commitment in both myeloid and lymphoid mouse progenitors. Blood 119: 2003–2012 [PubMed]
6. Holtschke T, Lohler J, Kanno Y, Fehr T, Giese N, et al. (1996) Immunodeficiency and chronic myelogenous leukemia-like syndrome in mice with a targeted mutation of the ICSBP gene. Cell 87: 307–317 [PubMed]
7. Schwieger M, Lohler J, Friel J, Scheller M, Horak I, et al. (2002) AML1-ETO inhibits maturation of multiple lymphohematopoietic lineages and induces myeloblast transformation in synergy with ICSBP deficiency. J Exp Med 196: 1227–1240 [PMC free article] [PubMed]
8. Turcotte K, Gauthier S, Tuite A, Mullick A, Malo D, et al. (2005) A mutation in the Icsbp1 gene causes susceptibility to infection and a chronic myeloid leukemia-like syndrome in BXH-2 mice. J Exp Med 201: 881–890 [PMC free article] [PubMed]
9. Diaz-Blanco E, Bruns I, Neumann F, Fischer JC, Graef T, et al. (2007) Molecular signature of CD34(+) hematopoietic stem and progenitor cells of patients with CML in chronic phase. Leukemia 21: 494–504 [PubMed]
10. Schmidt M, Hochhaus A, Nitsche A, Hehlmann R, Neubauer A (2001) Expression of nuclear transcription factor interferon consensus sequence binding protein in chronic myeloid leukemia correlates with pretreatment risk features and cytogenetic response to interferon-alpha. Blood 97: 3648–3650 [PubMed]
11. Lee KY, Geng H, Ng KM, Yu J, van Hasselt A, et al. . (2008) Epigenetic disruption of interferon-gamma response through silencing the tumor suppressor interferon regulatory factor 8 in nasopharyngeal, esophageal and multiple other carcinomas. Oncogene. [PubMed]
12. Silva FP, Swagemakers SM, Erpelinck-Verschueren C, Wouters BJ, Delwel R, et al. (2009) Gene expression profiling of minimally differentiated acute myeloid leukemia: M0 is a distinct entity subdivided by RUNX1 mutation status. Blood 114: 3001–3007 [PubMed]
13. Schmidt M, Nagel S, Proba J, Thiede C, Ritter M, et al. (1998) Lack of interferon consensus sequence binding protein (ICSBP) transcripts in human myeloid leukemias. Blood 91: 22–29 [PubMed]
14. Qian Z, Fernald AA, Godley LA, Larson RA, Le Beau MM (2002) Expression profiling of CD34+ hematopoietic stem/ progenitor cells reveals distinct subtypes of therapy-related acute myeloid leukemia. Proc Natl Acad Sci U S A 99: 14925–14930 [PubMed]
15. Stirewalt DL, Choi YE, Sharpless NE, Pogosova-Agadjanyan EL, Cronk MR, et al. . (2009) Decreased IRF8 expression found in aging hematopoietic progenitor/stem cells. Leukemia 23: 391–393. PMCID: PMC2640437. [PMC free article] [PubMed]
16. Anderson JE, Kopecky KJ, Willman CL, Head D, O'Donnell MR, et al. (2002) Outcome after induction chemotherapy for older patients with acute myeloid leukemia is not improved with mitoxantrone and etoposide compared to cytarabine and daunorubicin: a Southwest Oncology Group study. Blood 100: 3869–3876 [PubMed]
17. Petersdorf SH, Rankin C, Head DR, Terebelo HR, Willman CL, et al. (2007) Phase II evaluation of an intensified induction therapy with standard daunomycin and cytarabine followed by high dose cytarabine for adults with previously untreated acute myeloid leukemia: a Southwest Oncology Group study (SWOG-9500). Am J Hematol 82: 1056–1062 [PubMed]
18. Godwin JE, Kopecky KJ, Head DR, Willman CL, Leith CP, et al. (1998) A double-blind placebo-controlled trial of granulocyte colony-stimulating factor in elderly patients with previously untreated acute myeloid leukemia: a Southwest oncology group study (9031). Blood 91: 3607–3615 [PubMed]
19. List AF, Kopecky KJ, Willman CL, Head DR, Persons DL, et al. (2001) Benefit of cyclosporine modulation of drug resistance in patients with poor-risk acute myeloid leukemia: a Southwest Oncology Group study. Blood 98: 3212–3220 [PubMed]
20. Stirewalt DL, Kopecky KJ, Meshinchi S, Appelbaum FR, Slovak ML, et al. (2001) FLT3, RAS, and TP53 mutations in elderly patients with acute myeloid leukemia. Blood 97: 3589–3595 [PubMed]
21. Stirewalt DL, Kopecky KJ, Meshinchi S, Engel JH, Pogosova-Agadjanyan EL, et al. (2006) Size of FLT3 internal tandem duplication has prognostic significance in patients with acute myeloid leukemia. Blood 107: 3724–3726 [PubMed]
22. Ostronoff F, Othus M, Ho PA, Kutny M, Geraghty DE, et al. . (2012) Mutations in the DNMT3A exon 23 independently predict poor outcome in older patients with acute myeloid leukemia: a SWOG report. Leukemia. [PubMed]
23. Stirewalt DL, Meshinchi S, Kopecky KJ, Fan W, Pogosova-Agadjanyan EL, et al. (2008) Identification of genes with abnormal expression changes in acute myeloid leukemia. Genes Chromosomes Cancer 47: 8–20 [PubMed]
24. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25: 402–408 [PubMed]
25. Scheller M, Foerster J, Heyworth CM, Waring JF, Lohler J, et al. (1999) Altered development and cytokine responses of myeloid progenitors in the absence of transcription factor, interferon consensus sequence binding protein. Blood 94: 3764–3771 [PubMed]
26. Tamura T, Nagamura-Inoue T, Shmeltzer Z, Kuwata T, Ozato K (2000) ICSBP directs bipotential myeloid progenitor cells to differentiate into mature macrophages. Immunity 13: 155–165 [PubMed]
27. Tsujimura H, Tamura T, Gongora C, Aliberti J, Reis e Sousa C, et al. (2003) ICSBP/IRF-8 retrovirus transduction rescues dendritic cell development in vitro. Blood 101: 961–969 [PubMed]
28. Lucioni M, Novara F, Fiandrino G, Riboni R, Fanoni D, et al. (2011) Twenty-one cases of blastic plasmacytoid dendritic cell neoplasm: focus on biallelic locus 9p21.3 deletion. Blood 118: 4591–4594 [PubMed]
29. Yang D, Thangaraju M, Greeneltch K, Browning DD, Schoenlein PV, et al. (2007) Repression of IFN regulatory factor 8 by DNA methylation is a molecular determinant of apoptotic resistance and metastatic phenotype in metastatic tumor cells. Cancer Res 67: 3301–3309 [PubMed]

Articles from PLoS ONE are provided here courtesy of Public Library of Science