Molecular synergy underlies the co-occurrence paterns and phenotype of NPM1-mutant
acute myeloid leukemia.
Oliver M Dovey1, Jonathan L Cooper1, Annalisa Mupo1, Carolyn S Grove1,3,4, Claire Lynn5, Paul Green1, Nathalie Conte6, Robert M Andrews7, Kevin Shannon8, Tyler Jacks9, Roland Rad10, Mark Arends11, Penny Wright2,Allan Bradley1, Ignacio Varela12, and George S Vassiliou1-2†.
1
The Wellcome Trust Sanger Insttute, Wellcome Trust Genome Campus, Hinxton, Cambridge, UK;
2Department of Haematology, Cambridge University Hospitals NHS Trust, Cambridge, UK
3School of Pathology and Laboratory Medicine, University of Western Australia, Crawley, Western Australia, Australia 4
PathWest Division of Clinical Pathology, Queen Elizabeth II Medical Centre, Nedlands, Western Australia, Australia
5
Leukemia and Stem Cell Biology Group, Division of Cancer Studies, Department of Haematological Medicine, King's College London, Denmark Hill Campus, London SE5 9NU, UK
6
Mouse Informatcs, European Bioinformatcs Insttute (EMBL-EBI), Wellcome Trust Genome Campus, Hinxton, Cambridge, UK
7
Insttute of Translaton, Innovaton, Methodology and Engagement, Cardif University School of Medicine, Heath Park, Cardif, UK
8
Department of Pediatrics, University of California San Francisco, San Francisco, CA; Helen Diller Family Comprehensive Cancer Center, University of California San Francisco, San Francisco, California, USA
9
David H. Koch Insttute for Integratve Cancer Research, Massachusets Insttute of Technology, Cambridge, Massachusets 02139, USA; Howard Hughes Medical Insttute, Massachusets Insttute of Technology, Cambridge, Massachusets, USA 10
Department of Medicine II, Klinikum Rechts der Isar, Technische Universität München, 81675 Munich, Germany; German Cancer Consortum (DKTK), German Cancer Research Center (DKFZ), 69120 Heidelberg, Germany
11
Edinburgh Cancer Research Centre, University of Edinburgh, Western General Hospital, Edinburgh, UK
12
Insttuto de Biomedicina y Biotecnologaa de Cantabria, Santander 39011, Spain.
†Correspondence: gsv20@sanger.ac.uk
Running Title: Molecular synergy in NPM1 mutant AML
Include artcle word count here: 3992
Abstract Word Count: 248
Scientic Category: Myeloid Neoplasia 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Number of igures: 5
Number of references: 33
Key Points
1 or 2 key points no more than 140 characters
NPM1c and NRAS-G12D cooperate to cause high penetrance AML in mice.
Aggressive onset of mutant NPM1-FLT3 AML is underpinned by distnctve molecular and cellular synergies. 32 33 34 35 36 37 38
Abstract
Mutatons afectng NPM1 deine the commonest subgroup of acute myeloid leukemia
(AML). They frequently co-occur with mutatons of FLT3, usually internal tandem
duplicatons (ITD), and less commonly of NRAS or KRAS. Co-occurrence of mutant NPM1
with FLT3-ITD carries a signiicantly worse prognosis than the NPM1-RAS combinaton. To
understand the molecular basis of these observatons we compare the efects of the two
combinatons on hematopoiesis and leukemogenesis in knock-in mouse models. Early
efects of these mutatons on hematopoiesis show that compound Npm1
cA/+;Nras
G12D/+or
Npm1
cA;Flt3
ITDshare a number of features: Hox gene over-expression, enhanced
self-renewal, expansion of hematopoietc progenitors and a bias towards myeloid
diferentaton. The most notable diferences were that Npm1
cA;Flt3
ITDmutants, displayed
signiicantly higher peripheral leucocyte counts, early depleton of common lymphoid
progenitors and a monocytc bias compared to the granulocytc bias observed in Npm1
cA/+
;Nras
G12D/+mutants. Underlying this was a striking molecular synergy manifested as a
dramatcally altered gene expression proile in Npm1
cA;Flt3
ITD, but not Npm1
cA/+;Nras
G12D/+,
progenitors compared to wild type. Both compound models developed high penetrance
AML although latency in Npm1
cA/+;Nras
G12D/+mutants was signiicantly longer (median
survival 138 days post-pIpC in Npm1
cA/+;Nras
G12D/+vs 52.5 days in Npm1
cA;Flt3
ITDmice). During
AML evoluton, both models acquired additonal copies of the mutant Flt3 or Nras alleles,
but only Npm1
cA/+;Nras
G12D/+mice showed acquisiton of other mutatons observed in human
AML, including IDH1 R132Q. Our results show that molecular complementarity underlies the
frequent co-occurrence of mutant NPM1 and FLT3-ITD, and the poorer AML prognosis
associated with this mutaton combinaton compared to NPM1-NRAS.
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
Introducton
Advances in genomics have deined the somatc mutatonal landscape of acute myeloid leukemia (AML), leading to a detailed characterisaton of their prognostc signiicance and paterns of mutual co-occurrence or exclusivity.1, 2 Mutatons in NPM1, the gene for Nucleophosmin, characterise the most common subgroup of AML representng 25-30% of all cases, result in cytoplasmic dislocaton of the protein (NPM1c) and are mutually exclusive of leukemogenic fusion genes.1-3 As is ofen the case for fusion genes, progression to AML afer the acquisiton of mutant NPM1 is contngent upon the gain of additonal somatc mutatons such as those that actvate STAT and/or RAS signalling.4 For reasons that are not clear, this transforming step favours acquisiton of internal tandem duplicatons in FLT3 (FLT3-ITD) over other somatc mutatons with similar efects such as those involving NRAS or
KRAS.1-4 Furthermore, the NPM1c/FLT3-ITD combinaton is associated with a signiicantly worse prognosis compared to combinatons of NPM1c with mutant NRAS, KRAS or other mutatons.2 Whilst the adverse prognostc impact of NPM1/FLT3-ITD vs NPM1/RAS co-mutaton infuences clinical decisions in AML, its molecular basis and that of the frequent co-occurrence of NPM1c and
FLT3-ITD in AML are unknown. Here, in order to investgate these phenomena, we compare the
interacton of Npm1c with Flt3-ITD to its interacton with NrasG12D in knock-in mice. Individually,
knock-in models of NPM1c, FLT3-ITD and NRAS-G12D display enhanced myelopoiesis and progression to myeloproliferatve disorders or AML in a signiicant proporton of animals.5-7 Also, we and others have previously shown that Npm1c and Flt3-ITD synergise to drive rapid-onset AML, but the interacton between Npm1c and mutant NrasG12D has not, to our knowledge, been previously
investgated in knock-in mice.8 Our indings reveal that the combinaton of Npm1c and Flt3-ITD has an early profound efect on gene expression and hemopoiesis, whilst Npm1c and Nras-G12D display only modest molecular synergy and subtler cellular changes. Also, whilst both types of co-mutaton drove AML in the majority of mice, the leukemias in Npm1c;Flt3-ITD mice were more aggressive and undiferentated than those which developed in Npm1c;Nras-G12D animals. At the genomic level, there was frequent ampliicaton in both models of the mutant Flt3-ITD or Nras-G12D allele, however additonal somatc mutatons in AML driver genes (e.g. Idh1 and Ptpn11) were seen only in
Npm1c;Nras-G12D AMLs. Our indings propose that the molecular synergy between Npm1c and Flt3-ITD underpin the co-occurrence paterns, phenotype and prognosis of NPM1-mutant AML.
67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98
Materials and methods
Animal husbandry
Mx1-Cre+;Npm1fox-cA/+ were crossed with NrasLSL-G12D or Flt3ITD mice, to generate triple transgenic
animals (Mx1-Cre;Npm1fox-cA/+;NrasLSL-G12D/+ and Mx1-Cre;Npm1fox-cA/+;Flt3ITD/+).9 To actvate conditonal alleles (Npm1cA and NrasG12D) in approximately 12-14 week old Mx1-Cre;Npm1fox-cA/+;NrasLSL-G12D/+ mice,
Mx1-Cre was induced by intraperitoneal administraton of 5 doses of 200g pIpC over a 10 day
period. As described recently, Mx-1 Cre;Npm1Flox-cA/+;Flt3ITD/+ mutants do not require pIpC inducton of
Mx1-Cre and recombinaton of the Npm1fox-cA allele8. For pre-leukemic analyses Npm1cA/+;NrasG12D/+
were sacriiced 4-5 weeks post pIpC and Npm1cA/+;Flt3ITD/+ were sacriiced at 5 weeks of age.
Genotyping for mutant alleles was performed as previously.5, 10
Hematological measurements
Blood counts were performed on a VetABC analyzer (Horiba ABX).
Histopathology
Formalin ixed, parafn embedded (FFPE) sectons were stained with hematoxylin and eosin. Samples from leukemic mice were also stained with ant-CD3, ant-B220 and ant-myeloperoxidase, and detected using immunoperoxidase. All material was examined by two experienced histopathologists (P.W. and M.A.) blinded to mouse genotypes. Selected samples were also studied for total ERK1/2 (p44/42 MAPK, clone 137F5, Cell Signalling) and pERK1/2 (phosphor-p44/42 MAPK, clone 197G2, Cell Signalling).
Colony-forming assays and serial re-platng
Nucleated cells (3 x104) from bone marrow aspirates of mutant and wild-type mice were suspended in cytokine-containing methylcellulose-based media (M3434, Stem Cell Technologies) and plated in duplicate wells of 6-well plates. Colony-forming units (CFUs) were counted 7 days later. For serial re-platng, 3 x104 cells were re-seeded and colonies counted afer 7 days.
Flow cytometry and cell sortng
For fow cytometry, single cell suspensions of bone marrow cells or splenocytes were passed through a 0.4m nylon ilter and suspended in 0.85% NH4Cl for 5 minutes to lyse erythrocytes. Cells were then suspended in Hank's Balanced Salt Soluton (HBSS) supplemented with 2% FCS and 10 M HEPES. Progenitor populatons were deined and stained as described in supplementary methods. 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
Gated cellularity was calculated by multplying the percentage of gated cells by the total number of nucleated cells from bone marrow samples afer erythrocyte depleton.
Retroviral transducton of bone marrow progenitors
Lineage depleted bone marrow aspirates, isolated from wildtype and Flt3ITD/+ mice, were transduced
with MSCV-Hoxa9-GFP and/or MSCV-Nkx2-3-CFP retroviruses and expanded for 7 days in liquid culture (X-Vivo, Lonza, supplemented with 10ng/ml IL-3, 10ng/ml IL-6 and 50ng/ml SCF, Peprotech). CFP, GFP or double positve cells were FACS sorted and 2.5 x104 cells re-plated in semi-solid media as previously described. For cloning strategy, see supplemental methods.
Microarray and comparatve genomic hybridizaton analysis
Mouse gene expression proiles were generated using the Illumina MouseWG-6 v2 Expression BeadChip platorm (Illumina). DNA copy number variaton in leukemic samples was assessed using the Mouse Genome Comparatve Genomic Hybridizaton 244K Microarray (Agilent Technologies). Full details of analysis are provided in supplemental methods. For mouse gene expression proiling, n=4-10 (Lin-) or n=3-5 (MPP).
AML exome sequencing and mutaton calling
Whole exome sequencing (WES) of AML bone marrow and control C57BL/6N or 129Sv tail DNA was performed using the Agilent SureSelect Mouse Exon Kit (Agilent Technologies) and paired-end sequencing on a HiSeq2000 sequencer (Illumina). Validaton of mutatons was performed using MiSeq sequencing (Illumina) of amplicon libraries as described before (See Supplemental Methods Figure S1 and Supplemental Tables 6 and 7 for primer sequences).11 Full details of analysis are provided in supplemental methods.
Statstcs
Student t test or one-way analysis of variance (ANOVA, Bonferroni adjusted) were used for statstcal comparisons as appropriate and unless stated. Error bars represent standard error of the mean (SEM). Signiicant values are reported as: * P<0.05 Vs wildtype, ** P<0.01 Vs wildtype, *** P<0.001 Vs wildtype, P<0.05 Vs Flt3ITD/+, P<0.01 Vs Flt3ITD/+, P<0.001 Vs Flt3ITD/+
), P<0.05 Vs
NrasG12D/+
, P<0.01 Vs NrasG12D/+
, P<0.001 Vs NrasG12D/+, † P<0.05 Npm1cA/+; NrasG12D/+ Vs
Npm1cA/+; Flt3ITD/+, †† P<0.01 Npm1cA/+; NrasG12D/+ Vs Npm1cA/+; Flt3ITD/+, ††† P<0.001 Npm1cA/+; NrasG12D/ + Vs Npm1cA/+; Flt3ITD/+. 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157
Results
Mutant Npm1 co-operates with Nras-G12Dand Flt3-ITD to increase self-renewal of hematopoietc
progenitors and expand myelopoiesis
To understand the impact of the studied mutatons alone and in combinaton, we studied hemopoietc cell compartments of Npm1cA/+;NrasG12D/+, Npm1cA/+;Flt3ITD/+, Npm1cA/+; NrasG12D/+, Flt3ITD/+
and wild type (WT) control mice sacriiced 4-6 weeks afer actvaton of conditonal mutatons (Figure 1). Compared to Flt3ITD/+ singlemutants, Npm1cA/+; Flt3ITD/+ mice showed a marked increase of
peripheral WCCs (56±13.4 vs 6.5±0.5 x106 g/L, p<0.001) and spleen weights (0.63g vs 0.16g, p<0.001), but not of total bone marrow cellularity (Figure 1B). By contrast, both NrasG12D/+ and
Npm1cA/+;NrasG12D/+ mutants exhibited only subtle increases in spleen size (WT: 0.12g, NrasG12D/+:
0.18g, Npm1cA/+;NrasG12D/+:0.19g, p<0.01 and p<0.001 respectvely) and bone marrow cellularity (WT:
28.1±1.9 x 106, NrasG12D/+: 43.7±2.6 x 106 and Npm1cA/+;NrasG12D/+:41.3±3.2 x 106, p<0.01 compared to WT) (Figure 1B).
Expanded myelopoiesis and myeloproliferaton were previously documented in single NrasG12D/+ and
Flt3ITD/+ mutant mice.6, 7 The additon of mutant Npm1 augmented these phenotypes (Supplemental Figure S1A). In partcular, total Mac-1+ splenocytes increased in number (27%-50% for NrasG12D/+; and
from 57-73% for Flt3ITD/+). Notably, these cells were predominantly granulocytc (Mac-1+/Gr-1+) in
Npm1cA/+; NrasG12D/+ mice in contrast to cells from Npm1cA/+; Flt3ITD/+ mice, which were predominantly
monocytc Mac-1+/Gr-1- (Supplemental Figure S1A).
NrasG12D/+ mice have been shown to have increased hematopoietc stem and progenitor cell
(HSCP)numbers, due to increased proliferaton and self-renewal of the HSC and MPP compartments.12, 13 Our results conirm these data demonstratng signiicant increases in total myeloid progenitors, GMPs and CMPs, as well as the total number of early progenitors (LSK, and MPP) in both Npm1cA/+; NrasG12D/+ and NrasG12D/+ bone marrow cells when compared to WT controls
(Figure 1C and Supplemental Figure S2A). Our data also reveal that NrasG12D/+ stem and progenitor
cell compositon is largely unaltered by the additon of mutant NPM1. Also in agreement with previous studies, hematopoiesis in Flt3ITD/+ mice was characterised by increased numbers of total
myeloid progenitors (LK p<0.05 and GMPs p<0.01) and early progenitor populatons (LSK, MPP and LMPP, p<0.01, p<0.01 and p<0.05 respectvely) (Figure 1C and Supplemental Figure S2A). 14, 15 Of note, there was a substantal decrease in the size of the common lymphoid progenitor (CLP) populaton in Flt3ITD/+ and Npm1cA/+;Flt3ITD/+ mice (Figure 1C) but not in single or compound NrasG12D/+
mutants. This was in part due to the reducton in Il-7Rα-positve cells (Figure S2B). Npm1cA/+;Flt3ITD/+
158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
mice also exhibited robust increases in numbers of LK, GMP, LSK, MPP and LMPP populatons (over what was observed with Flt3ITD/+) when compared to WT. In direct comparison with Flt3ITD/+ mutants,
numbers of CMP and MEP progenitors in Npm1cA/+; Flt3ITD/+ mice were reduced (from 55x103 to 16x103, p<0.05 and from 61x103 to 17x103, p<0.05), yet GMPs (proposed as direct descendants of CMPs 16) remain expanded when compared to Npm1cA/+;NrasG12D/+ mutants. This demonstrates that
Flt3ITD/+ mutant myelopoiesis is dramatcally altered by the additon of Npm1cA/+. Also, when
compared to Npm1cA/+;NrasG12D/+ mice, Npm1cA/+;Flt3ITD/+ mice showed an increase in LMPPs,
reducton in lymphoid progenitors (CLP) and increase in GMPs (Figure 1E).
In order to assess the efects on the earliest detectable hematopoietc stem cell compartment (HSC) we opted to perform E-SLAM staining (CD45+/EPCR+/CD48-/CD150+).17 Importantly, and unlike many other HSC FACS strategies, this does not rely on cell surface expression of FLT3, and reveals the percentage of E-SLAM detectable HSCs is decreased in Npm1cA/+;NrasG12D/+ mice further so in Npm1cA/ +;Flt3ITD/+mutants (Figure 1D). Finally, using serial re-platng of bone marrow cells in semi-solid media
we show that Npm1cA/+ co-mutaton markedly increased self-renewal of Flt3ITD/+ cells (as shown
previously 8) and also of NrasG12D/+ (Figure 1F).
The Npm1cA/+ transcriptonal signature persists in double mutant hemopoietc progenitors
To examine their combined efects on transcripton we performed comparatve global gene expression proiling of lineage negatve (Lin-) bone marrow cells using microarrays. Npm1cA/+;
NrasG12D/+ and Npm1cA/+;Flt3ITD/+ cells displayed a dramatcally altered gene expression proile
compared to single NrasG12D/+ or Flt3ITD/+ mutants (Figure 2A and Supplemental Figure S3B).
Previously, we showed that mouse Npm1cA/+ Lin- cells overexpressed several homeobox (Hox) genes (in partcular overexpression of Hoxa5, Hoxa7, Hoxa9 and two other homeobox genes, Hopx and
Nkx2-3).18 Here, we show that this signature, absent from NrasG12D/+ or Flt3ITD/+ singular mutant mice,
persists in compound Npm1cA/+;NrasG12D/+ and Npm1cA/+;Flt3ITD/+ Lin- progenitors. (Figure 2A, Supplemental Figure S3A-C). Gene Set Enrichment Analysis (GSEA) of Npm1cA/+ single and compound
mutant cell expression proiles, showed signiicant enrichment for genes up-regulated in NPM1-mutant and MLL-fusion gene positve human leukemias (Figure 2A).
Overexpression of the homeobox gene NKX2.3 in NPM1-mutant AML
Using the human TCGA AML dataset, we compared gene expression proiles of NPM1 mutant (NPM1c+ve) to NPM1 wildtype (NPM1wt) AML.1 In agreement with previously published analyses, both 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220
HOXA and HOXB genes were signiicantly overexpressed in NPM1c+ve AML (Figure 2B).19 We also noted that another homeobox gene, NKX2-3, was also overexpressed in keeping with our indings in
Npm1cA/+ mice (Figure 2A). Recently, NKX2-3 overexpression was shown to be the most efectve
discriminant of MLL-MLLT4 (MLL-AF6) driven AML to AML driven by other MLL-fusion genes. 20 Whilst overexpression of Hox genes such as Hoxa9 has been shown to impart increased self-renewal and proliferaton of hematopoietc progenitors, the efects of Nkx2-3 overexpression are unknown. 21 To study this we performed retroviral gene transfer of fuorescently tagged Nkx2-3-CFP and Hoxa9-GFP into wildtype and Flt3ITD/+ Lin- bone marrow cells. Cells were subsequently sorted and plated in semi-solid methylcellulose for colony formaton assays (Figure 2Ci). We ind that overexpression of
Nkx2-3 increases clonogenic potental, albeit to a lesser extent compared to Hoxa9 overexpression,
in both wildtype and Flt3ITD/+ progenitors. Notably, this is not augmented in combined transfected
cells. (Figure 2Cii).
In order to mitgate the impact of the studied driver mutatons on cell surface phenotypes, we performed transcriptome analysis on a homogeneous populaton of puriied LSK-multpotent progenitor cells (MPPs, cell surface proile: Lin-/CD34+/Flt3+/CD48+/CD150-) (Figure 2D). Transcripton proiles of MPPs from single NrasG12D/+ or Flt3ITD/+ and the respectve Npm1cA/+ compound mutant
MPPs revealed distnct transcriptonal changes. In partcular, compared to WT, both NrasG12D/+ and
Npm1cA/+;NrasG12D/+ MPPs displayed similarly small numbers of diferentally expressed genes yet only
~20% of these were shared (Figure 2Di). Also, gene/pathway enrichment analyses did not uncover signiicant overlap with a pre-established expression signature (data not shown). In contrast, the “additon” of Npm1cA/+ to Flt3ITD/+ in MPPs led to diferental expression of a large number of
additonal genes, whilst also retaining most of the transcriptonal changes atributable to Flt3ITD/+
(Figure 2Dii). This demonstrates the powerful synergy between Npm1cA/+ and Flt3ITD/+ (Figure 2Dii,
Tables S2a-b). Pathway analysis of genes diferentally expressed in Npm1cA/+; Flt3ITD/+ MPPs revealed
enrichment of genes in the JAK-STAT pathway (Supplemental Figure 2E, Supplemental Tables S4a-b), including the negatve regulators Cish and Socs2 (Figure 2F). A number genes, encoding proteins involved in MAPK signalling are also deregulated (i.e. down-regulaton of the MAPK pathway inhibitor, Dusp6 and up-regulaton of actvators, Rasgrp1, Rasgrf2 and RaslIIb) (Figure 2F). Other notable dysregulated genes included those involved in chromatn organisaton (down-regulated,
Hdac10, Hdac11, Cbx7, Fbxl10/Kdm2b, Chd3, Satb1 and H1F0) and hematopoietc lineage or myeloid
cell diferentaton (Bcl6, Bmp1, Lmo2, Ldb1 all down-regulated with Cd74 and Thy1 up-regulated) (Figure 2F, Supplemental Figure 3D). Of partcular note is the down regulaton of two positve regulators of murine lymphoid hemopoiesis, Bcl11a and Kdm2a/Fbxl10 (knock-out mice of either gene are devoid of detectable CLPs22, 23). Many of the genes in our Npm1cA/+; Flt3ITD/+ data set are also
221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254
present as deregulated in a recently published Tet2-/-;Flt3ITD/+ mouse model that also develops AML
(76/418 genes, Supplemental Figure 3F and Supplemental Table 6,) which serves to verify our dataset but also reveals a distnguishing expression signature of FLT3-ITD which includes Socs2, Id1,
Csfr3r and Bcl11a. In contrast a lack of correlaton between deregulated gene sets of Npm1cA/+;Flt3ITD/ + and Npm1cA/+;NrasG12D/+ MPPs (Supplemental Figure S3E) emphasises the molecular distncton
between these compound mutants. 24
Notably, Hox gene expression was not signiicantly altered in MPP populatons from any of the
Npm1cA/+ models when compared to wildtype or to single NrasG12D/+ and Flt3ITD/+ mutants (Figure 2E
and Figure S3C). These results are in agreement with observatons that Hox gene expression in human NPM1c+ve AML is comparable to that seen in normal HSCs and myeloid progenitors.19 However, although we observe expanded myeloid progenitor populatons, these data propose that the observed patern of homeobox gene dysregulaton is a direct molecular consequence of NPM1c rather than a change in cellular compositon.
Npm1cA/+ and NrasG12D collaborate to promote high penetrance AML
To understand the leukemogenic potental of combined Npm1cA/+ and NrasG12D mutatons, we aged
cohorts of Npm1cA/+;NrasG12D/+ mice along with Npm1cA/+;Flt3ITD/+, single Npm1cA/+, NrasG12D and Flt3ITD/+
mutant and wildtype control mice. Compound Npm1cA/+;NrasG12D/+ and Npm1cA/+;Flt3ITD/+ mice had
signiicantly reduced survival (median 138 and 52.5 days respectvely) when compared to wildtype (618 days), Npm1cA/+ (427 days), NrasG12D/+ (315 days) and Flt3ITD/+ (also 315 days) (Figure 3A,
Supplemental Figure S4A). No diference in the survival of NrasG12D/+ and Flt3ITD/+ mutant mice was
observed (p value = 0.85, see Supplemental Figure S4A for all comparisons). Moribund Npm1cA/ +;NrasG12D/+ and Npm1cA/+;Flt3ITD/+ mice displayed increased spleen and liver weights compared to
moribund mice of other genotypes (Figure 3B). At tme of sacriice they also had signiicantly increased blood leukocyte (32.6±12, NrasG12D/+ compared to 359±62 x106/L, Npm1cA/+;NrasG12D/+; and
151±34, Flt3ITD/+ compared 250±33 x106/L, Npm1cA/+; Flt3ITD/+) and reduced platelet counts (1046±227,
NrasG12D/+ compared to 504.9±209 x106/L, Npm1cA/+; NrasG12D/+; and 607±99, Flt3ITD/+ compared to
225±25.8 x106/L, Npm1cA/+; Flt3ITD/+). Iniltraton of spleen tssue with myeloid cells was conirmed by
FACS (Mac-1/Gr-1) (Supplemental Figure S4B). Independent histopathological analysis of formalin ixed parafn embedded tssues from moribund mice (Figure 3C) revealed an increase in AML incidence from 41% (Flt3ITD/+) to 100% in Npm1cA/+;Flt3ITD/+ samples and from 13% (NrasG12D/+) to 85%
in Npm1cA/+;NrasG12D/+ samples (45% AML with maturaton, AML+ and 40% AML without maturaton, AML- as deined by Bethesda classiicaton) (Figure 3C).
255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
Additonal somatc mutatons are required for progression to AML in Npm1cA/+
; NrasG12D/+ mice. Npm1cA/+;Flt3ITD/+ mice succumb to AML very rapidly, compared to Npm1cA/+ and Npm1cA/+;NrasG12D/ +mice. We hypothesised that the slower onset of AML in the later two genotypes may be due to the
requirement for additonal cooperatng somatc mutatons. To test this, we performed whole exome sequencing and array comparatve hybridisaton (aCGH) of AMLs from Npm1cA/+, Npm1cA/+;Flt3ITD/+and
Npm1cA/+;NrasG12D/+ mice. We irst conirmed the frequent development of loss-of-heterozygosity
(LOH) at the Flt3 locus in Npm1cA/+;Flt3ITD/+ AMLs 8, 25 and used quanttatve amplicon sequencing (MiSeq; Methods, Supplemental Methods Figure S1) to quantfy the Flt3ITD variant allele fractons
(VAFs), which were greater than 0.5 in 5/5 (range: 0.55-0.95, Figure 4Ai). Results of aCGH showed that LOH was copy-neutral (i.e. acquired uniparental disomy, Supplemental Figure 4Aii), with duplicaton of the Flt3ITD allele. Using another quanttatve amplicon sequencing assay we also show
that recombinaton of the Npm1fox-cA allele was complete or near-complete (Figure S5A).
Interestngly, aCGH of Npm1cA/+;NrasG12D/+ samples revealed ampliicaton of chr3 in 5/10 samples
tested (Figure 4Bi). This was exclusively seen in Npm1cA/+;NrasG12D/+ AMLs and mapped to a minimally
ampliied region (chr3: 102743581-103470550) which contained the genes Nr1h5, Sike1, Csde1,
Ampd1, Dennd2c, Bcas, Trim33 and Nras (Supplementary Table S10). This was conirmed by
amplicon speciic MiSeq PCR of the NrasG12D allele and demonstrates gains of mutant Nrasin these
samples. This assay also identied gains of mutant Nras in 3 of 5 copy neutral samples by aCGH at the Nras locus (and 3 out of 4 further samples not assayed by aCGH). In summary, increased NrasG12D
dosage was detected in 11/14 Npm1cA/+;NrasG12D/+ AMLs (Figure 4Bii). Staining of FFPE tssues from
Npm1cA/+;NrasG12D/+ AMLs for pERK1/2 (actvated downstream of mutant RAS), demonstrated the
level of RAS pathway actvaton approximately correlated with NrasG12D gene dosage (Figure 4C).
Whole exome sequencing of AML samples revealed the average cumulatve number of single nucleotde variants (SNVs) and small insertons or deletons (Indels) per AML sample was positvely correlated to median survival rates. That is, Npm1cA/+ 6.8±0.9, Npm1cA/+; NrasG12D/+ 3.3±0.5 and
Npm1cA/+; Flt3ITD/+2.6±0.7 (mean number of mutatons ± S.E.M.) (Figure 5A). Given the short latency
to AML, it is unsurprising that Npm1cA/+;Flt3ITD/+ AMLs exhibit the fewest mutatons. Moreover,
Npm1cA/+ AMLs spontaneously acquire mutatons in genes involved in RAS signalling (NrasQ61H, CblS374F,
Ptpn11S502L and Nf1W1260*/R683*) conirming this genetc interacton. Likewise, detecton of a
spontaneous tyrosine kinase domain mutaton in Flt3, (Flt3D842G) further conirms the importance of
mutant FLT3 in the progression of mutant NPM1 AML (Figure 5B, Supplemental Table 9). Interestngly, another spontaneously acquired mutaton, in a single Npm1cA/+;NrasG12D/+ AML, occurs
287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319
in the Idh1 gene (R132Q) and although R132H/R132C are the commonly detected mutatons in human AML,1 the R132Q mutaton has been demonstrated as being an equivalent pathogenic variant in chondrosarcoma.26 aCGH also reveals the presence of complete or partal gain of chr7 in 7/8 Npm1cA/+ and 4/9 Npm1cA/+; NrasG12D/+ AMLs (Figure 5C and Supplemental Figure S5B). A minimally
ampliied region contains a number of genes previously implicated in leukemic transformaton such as Nup98,Wee1 and Eed (Supplemental Figure S5C).18, 28, 29 The ampliied region of murine chr7 is
syntenic to human chr11 (Supplemental Table S11). Trisomy 11 has been reported in human AML and MDS, ofen in concert with MLL-PTD mutatons.30, 31 Single copy loss of chromatn modifying enzymes commonly found in NK-AML, WT1, Asxl1 and Dnmt3b occurs simultaneously in one
Npm1cA/+ sample (Supplemental Figure S5D) and a focal deleton of Ezh2 is detected in a single
Npm1cA/+;NrasG12D/+ AMLs (Figure 5c). Other signiicant chromosomal gains include a region of chr6 in
RN8 (which contains the oncogene Ret as well as Kras) and two instances of chr15 ampliicaton, RN3 and RN6 which contain the Ghr gene, a common inserton site in our Npm1cA/+ transposon insertonal
mutagenesis screen (Figure 5C, Supplemental Figure S6 and Supplemental Table 10 for full results). 18
Discussion
Whilst the mutatonal drivers of AML and their paterns of co-occurrence are now well understood, the molecular basis for the frequency and prognostc impact of these paterns remains unknown. Of partcular clinical relevance are the co-occurrence paterns of mutant NPM1, the equal most common mutaton type in human AML.1,2 Co-mutaton of NPM1 with FLT3-ITD is both signiicantly more frequent and carries a worse prognosis than co-mutaton of NPM1 with NRAS or KRAS. To understand the basis of this observaton we investgated the interactons of these mutatons in bespoke experimental models (Figure 1A). We irst looked at the short-term impact of these mutatons on hemopoiesis in young mice and conirmed that single Npm1cA/+ mutant mice have
normal bone marrow cellularity, white blood cell counts (WCC) and splenic weight.18 Also, as described before, single Flt3ITD/+ and NrasG12D/+ had moderate but signiicant increases in splenic size,
whilst NrasG12D/+ also had raised WCC and bone marrow cellularity.6,7 However, whilst the introducton of Npm1cA/+ into the NrasG12D/+ background did not alter these parameters signiicantly,
the Npm1cA/+;Flt3ITD/+ co-mutaton led to a dramatc rise in white cell count and splenic size (Figure
1B). At the cellular level, the Npm1cA/+; NrasG12D/+ combinaton did not lead to signiicant changes on
the size of progenitor and stem cell compartments when compared to NrasG12D/+ alone. In sharp
contrast, when compared to Flt3ITD/+ single mutants, Npm1cA/+;Flt3ITD/+ mice displayed reductons in
320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
CMP and MEP and increases in LSK progenitors. Furthermore, Npm1cA/+;Flt3ITD/+ mice showed a
profound reducton in HSCs (Figure 1 C–E).
This diferental impact of Npm1cA/+ on Flt3ITD/+ versus NrasG12D/+ was refected in the marked
diferences in gene expression proiles (GEP) between these two types of double-mutant mice. In fact the Npm1cA/+;NrasG12D/+ model displayed only minimal diferences to single NrasG12D/+, whilst
Npm1cA/+;Flt3ITD/+ lin- progenitors had profoundly diferent GEPs to Flt3ITD/+ refectng the dramatc
impact of this combinaton on gene expression and in turn on progenitor cell fate. In order to discern the efects on transcripton from changes in cellular compositon, we studied gene expression in puriied MPPs. Interestngly, whilst the impact of “adding” Npm1cA/+ was much more dramatc with
Flt3ITD/+ than with NrasG12D/+ (Supplemental Figure S3), the characteristc Hox gene signature of
Npm1cA/+ was not detectable in MPPs (Figure 2D-F). By contrast, other diferences in MPP gene
expression between Npm1cA/+;NrasG12D/+ and Npm1cA/+;Flt3ITD/+ may have a role in the observed
cellular phenotypes, for example Kdm2b, a critcal gene for lymphopoiesis (is down-regulated in
Npm1cA/+;Flt3ITD/+ (Figure 2F and Supplemental Figure S3D), which displays a signiicant reducton in
CLPs.22 Our transcriptome analysis of mouse progenitors and human AML reveals overexpression of
Nkx2-3 to be a distnguishing molecular feature of mutant NPM1. NKX2-3 has been described as a
prominently expressed and a novel distnguishing marker of MLL-AF6 and MLL-ENL from other of forms of MLL leukemia (fusion or PTD). 20, 32 Here we report this phenomenon in human NPM1c+ve NK-AML. A full appreciaton of Nkx2-3 over-expression and how it may contribute to the efects on leukemogenesis has yet to be discerned and may warrant closer inspecton, especially as it is suggestve of a common mechanism of deregulaton of homeobox gene expression in these leukemia sub-types.
To study how these diferences impact on leukemogenesis we aged double mutant mice and report that, like Npm1cA/+;Flt3ITD/+ animals (Mupo et al8), Npm1cA/+;NrasG12D/+ mice also develop highly
penetrant AML. However, AML latency was markedly shorter in Npm1cA/+;Flt3ITD/+ (median, 52.5 days)
than in Npm1cA/+; NrasG12D/+ (median, 138 days) mice. Interestngly single mutant Flt3ITD/+ and NrasG12D/ + mice did not display a diferent survival (Figure 3A) indicatng that the interacton with Npm1cA was
central to this diference. Also, Npm1cA/+;Flt3ITD/+ AMLs lacked myeloid diferentaton, which was
frequently seen in Npm1cA/+;NrasG12D/+ leukemias. To understand the genetc events involved in
leukemic progression, we performed copy number analysis and exome sequencing of Npm1cA/ +;Flt3ITD/+ and Npm1cA/+;NrasG12D/+ AMLs. We found that the commonest somatc event during
progression was an increase in mutant allele burden (NrasG12D/+ or Flt3ITD/+), through copy neutral LOH
353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384
(6/6 Npm1cA/+;Flt3ITD/+ and 3/8 Npm1cA/+;NrasG12D/+ AMLs) or copy number gain (5/8 Npm1cA/+;NrasG12D/+
AMLs).
Using faithful mouse models, our study provides strong and comparatve evidence that the co-occurrence of mutant NPM1 with FLT3-ITD is more formidable in AML potentality, led by strikingly diferent molecular and cellular consequences, compared to its co-occurrence with mutant NRAS. This is a very plausible explanaton for the frequent co-occurrence and worse prognosis of double mutant NPM1c/ FLT3-ITD AML.
Acknowledgements
We wish to thank Prof. Gary Gilliland for the FLt3ITD mouse. OMD, JLC and GSV are funded by a Wellcome Trust Senior
Fellowship in Clinical Science (WT095663MA). AM is funded by the Kay Kendall Leukaemia Fund project grant. CG was funded by a Bloodwise Clinical Research Training Fellowship. IV is funded by Spanish Ministerio de Economaa y
Compettvidad subprograma Ramón y Cajal. We thank Servicio Santander Supercomputación for their support. GSV is a consultant for and holds stock in Kymab Ltd, and receives an educatonal grant from Celgene.
Authorship
Contributon: O.M.D., J.L.C., A.M., C.S.G., C.L., P.G. and G.S.V. performed mouse experiments. O.M.D. and G.S.V. analyzed results; P.W. and M.A. performed histopathological analysis of mouse samples; O.M.D., N.C. and R.M.A. performed microarray analysis; I.V. performed analysis of next generaton sequencing; O.M.D. and G.S.V. designed the study. O.M.D. and G.S.V. wrote the paper with the help of K.S., T.J., R.R., P.W., M.A. and A.B.
Confict of interest disclosure: GSV is a consultant for and holds stock in Kymab Ltd, and receives an educatonal grant from Celgene. All other authors declare no competng inancial interests.
Correspondence: George Vassiliou, 1
The Wellcome Trust Sanger Insttute, Wellcome Trust Genome Campus, Hinxton, Cambridge, CB10 1SA, UK; e-mail: gsv20@sanger.ac.uk.
References
1. Network TCGAR. Genomic and epigenomic landscapes of adult de novo acute myeloid leukemia. N Engl J Med. 2013; 368(22):2059-74.
2. Papaemmanuil E, Gerstung M, Bullinger L, et al. Genomic Classiicaton and Prognosis in Acute Myeloid Leukemia.
N Engl J Med. 2016; 374(23):2209-21.
3. Falini B, Mecucci C, Tiacci E, et al. Cytoplasmic nucleophosmin in acute myelogenous leukemia with a normal karyotype. N Engl J Med.
385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417
4. Welch JS, Ley TJ, Link DC, et al. The origin and evoluton of mutatons in acute myeloid leukemia. Cell. 2012; 150(2):264-78.
5. Vassiliou GS, Cooper JL, Rad R, et al. Mutant nucleophosmin and cooperatng pathways drive leukemia initaton and progression in mice. Nat Genet.
6. Lee BH, Tothova Z, Levine RL, et al. FLT3 mutatons confer enhanced proliferaton and survival propertes to multpotent progenitors in a murine model of chronic myelomonocytc leukemia. Cancer Cell. 2007; 12(4):367-80. 7. Li Q, Haigis KM, McDaniel A, et al. Hematopoiesis and leukemogenesis in mice expressing oncogenic NrasG12D
from the endogenous locus. Blood. 2011; 117(6):2022-32.
8. Mupo A, Celani L, Dovey O, et al. A powerful molecular synergy between mutant Nucleophosmin and Flt3-ITD drives acute myeloid leukemia in mice. Leukemia. 2013;27(9):1917-20.
9. Kuhn R, Schwenk F, Aguet M, Rajewsky K. Inducible gene targetng in mice. Science. 1995;269(5229):1427-9. 10. Haigis KM, Kendall KR, Wang Y, et al. Diferental efects of oncogenic K-Ras and N-Ras on proliferaton,
diferentaton and tumor progression in the colon. Nat Genet. 2008;40(5):600-8.
11. McKerrell T, Park N, Moreno T, et al. Leukemia-associated somatc mutatons drive distnct paterns of age-related clonal hemopoiesis. Cell Rep. 2015;10(8):1239-45.
12. Wang J, Kong G, Liu Y, et al. Nras(G12D/+) promotes leukemogenesis by aberrantly regulatng hematopoietc stem cell functons. Blood. 2013;121(26):5203-5207.
13. Li Q, Bohin N, Wen T, et al. Oncogenic Nras has bimodal efects on stem cells that sustainably increase compettveness. Nature . 2013 Dec 5;504(7478):143-147.
14. Shih AH, Jiang Y, Meydan C, et al. Mutatonal cooperatvity linked to combinatorial epigenetc gain of functon in acute myeloid leukemia. Cancer Cell. 2015;27(4):502-515.
15. Mead AJ, Kharazi S, Atkinson D, et al. FLT3-ITDs instruct a myeloid diferentaton and transformaton bias in lymphomyeloid multpotent progenitors. Cell Rep. 2013 Jun 27;3(6):1766-1776.
16. Akashi K, Traver D, Miyamoto T, Weissman IL. A clonogenic common myeloid progenitor that gives rise to all myeloid lineages. Nature. 2000 Mar 9; 404(6774):193-197.
17. Kent DG, Copley MR, Benz C, et al. Prospectve isolaton and molecular characterizaton of hematopoietc stem cells. Blood. 2009 Jun 18;113(25):6342-50
18. Vassiliou GS, Cooper JL, Rad R, et al. Mutant nucleophosmin and cooperatng pathways drive leukemia initaton and progression in mice. Nat Genet. 2011;43(5):470-475.
19. Spencer DH, Yound MA, Lamprecht TL, et al. Epigenomic analysis of the HOX gene loci reveals mechanisms that may control canonical expression paterns in AML and normal hematopoietc cells. Leukemia. 2015;29(6):1279-1289.
20. Lavalee VP, Baccelli I, Krosl J, et al. The transcriptomic landscape and directed chemical interrogaton of MLL-rearranged acute myeloid leukemias. Nat Genet. 2015;47(9):1030-1037.
21. Thorsteinsdotr U, Mamo A, Kroon E, et al. Overexpression of the myeloid leukemia-associated Hoxa9 gene in bone marrow cells induces stem cell expansion. Blood. 2002;99(1):121-129.
22. Andricovich J, Kai Y, Peng W, et al. Histone demethylase KDM2B regulates lineage commitment in normal and malignant hematopoiesis. J Clin Invest. 2016; 126(3):905-920.
23. Yu Y, Wang J, Khaled W, et al. Bcl11a is essental for lymphoid development and negatvely regulates p53. J Exp
Med. 2012; 209(13):2467-2483. 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
24. Shih AH, Jiang Y, Meydan C, et al. Mutatonal cooperatvity linked to combinatorial epigenetc gain of functon in acute myeloid leukemia. Cancer Cell. 2015 Apr 13;27 (4): 502-512.
25. Strewalt DL, Pogosova-Agadjanyan EL, Tsuchiya K, et al. Copy-neutral loss of heterozygosity is prevalent and a late event in the pathogenesis of FLT3/ITD AML. Blood Cancer J. 2014;4:e208
26. Hirata M, Sasaki M, Cairns RA, et al. Mutant IDH is sufcient to initate enchondromatosis in mice. Proc Natl Acad
Sci. 2015;112(9):2829-2834.
27. Weisberg E, Nonami A, Chen Z, et al. Identicaton of Wee1 as a novel therapeutc target for mutant RAS-driven acute leukemia and other malignancies. Leukemia. 2015 Jan;29(1):27-37.
28. Shi J, Wang E, Zuber J, et al. The Polycomb complex PRC2 supports aberrant self-renewal in a mouse model of MLL-AF9;Nras(G12D) acute myeloid leukemia. Oncogene . 2013;32(7):930-938.
29. Danis E, Yamauchi T, Echanique K, et al. Inactvaton of Eed impedes MLL-AF9-mediated leukemogenesis through Cdkn2a-dependent and Cdkn2a-independent mechanisms in a murine model. Exp Hematol. 2015;43(11):930-935. 30. Abdel-Wahab O, Levine R. The spliceosome as an indicted conspirator in myeloid malignancies. Cancer Cell.
2011;20(4):420-423.
31. Wang SA, Jabbar K, Lu G, et al. Trisomy 11 in myelodysplastc syndromes deines a unique group of disease with aggressive clinicopathologic features. Leukemia. 2010 Jan 24(4); 740-747
32. Garcia-Cuellar MP, Butner C, Bartenhagen C, et al. Leukemogenic MLL-ENL Fusions Induce Alternatve Chromatn States to Drive a Functonally Dichotomous Group of Target Genes. Cell Rep. 2016;15(2):310-322.
33. Dovey OM, Chen B, Mupo A, et al. Identicaton of a germline F692L drug resistance variant in cis with Flt3-ITD in knock-in mice. Haematologica. 2016;101(8):e328-331.
Figure Legends
Figure 1. Mutant Npm1 co-operates with Nras-G12D and Flt3-ITD to enhance myeloid diferentaton and enhance progenitor self-renewal.
(A) Schema for Mx-1 Cre, Npm1fox-cA, NrasLSL-G12D and Flt3ITD inter-crosses. (B) NrasG12D/+ mice show a
subtle and Npm1cA/+; Flt3ITD/+ mice a marked increase in white cell count (WCC), compared to
wildtype. Splenic sizes were signiicantly increased in all mutant genotypes except Npm1cA/+, with
Npm1cA/+; Flt3ITD/+ showing the most striking phenotype. Bone marrow cellularity was increased only
in the presence of the NrasG12D/+ allele. (C) FACS analysis at 4-5 weeks afer mutaton inducton.
Gatng strategies depicted are from wildtype mice. Signiicant diferences in the stem and progenitor cell compartments of NrasG12D/+ and Flt3ITD/+, but not Npm1cA/+ single mutant mice, as previously
reported. In double mutant mice, the Npm1cA/+; NrasG12D/+ combinaton was not signiicantly diferent
to NrasG12D/+, in contrast to Npm1cA/+;Flt3ITD/+ which was markedly diferent to both Flt3ITD/+ and
Npm1cA/+ single mutants. (D) Using a cell surface phenotype independent of FLT3 staining, we found
that CD45+/EPCR+/CD150+/CD48- HSCs were reduced slightly in Npm1cA/+;NrasG12D/+ and markedly in
Npm1cA/+;Flt3ITD/+mice. (E) Summary of hematopoietc efects of Npm1cA/+;NrasG12D/+ and Npm1cA/ +;Flt3ITD/+ double mutatons in mice. LK, Lin-/Kit+; LSK, Lin-/Sca-1+/Kit+; CMP, common myeloid progenitor; MEP, megakaryocyte-erythroid progenitor; GMP, granulocyte-monocyte progenitor; 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496
MPP, mult-potent progenitor; LMPP, lymphoid primed mult-potent progenitor; CLP, common lymphoid progenitor and HSC, hematopoietc stem cell. (F) Single Npm1cA/+ and double Npm1cA/ +;NrasG12D/+ or Npm1cA/+;Flt3ITD/+ mutant hematopoietc progenitors show increased self-renewal
potental in whole bone marrow serial replatng assays Mean ± SEM (n=4-8).
Figure 2. Impact of Npm1cA/+
on the transcriptomes of NrasG12D/+
and Flt3ITD/+
mutant hematopoietc progenitors.
(A) Overlap of diferentally expressed mRNAs reveals that Npm1cA/+ has a dramatc impact on
Lin-progenitor gene expression proiles when combined with Flt3ITD/+, but only a modest impact when
combined with NrasG12D/+. Nonetheless, the characteristc hallmarks of Npm1cA/+ are retained in both
double mutant progenitors, namely overexpression of Hoxa genes and of the homeobox genes Hopx and Nkx2-3 (also seen in single Npm1cA/+ progenitors). Gene Set Enrichment Analysis reveals
enrichment of diferentally expressed genes from these models in human AMLs harboring mutant
NPM1 or MLL gene fusions (B) Comparison of human NPM1-mutant (NPM1c)versus NPM1-wildtype
(NPM1WT) normal karyotype AML (NK-AML) also shows marked overexpression of HOXA and HOXB
genes, as well as of NKX2.3 raising the possibility that the later may mediate some of the efect of
NPM1c. (C) Efects of Nkx2-3 and Hoxa9 over-expression on mouse hematopoietc progenitors. (i) Lin -bone marrow progenitors from wildtype and Flt3ITD/+ mice were transduced with MSCV-Nkx2.3-CFP
and/or MSCV-Hoxa9-GFP constructs, maintained in liquid culture for 7 days, FACS sorted for CFP and GFP single and for double transfected cells and plated in semi-solid media. (ii) Colony assays of 2500 transduced cells show that both MSCV-Hoxa9 and MSCV-Nkx2-3 conferred an increase in self-renewal of both wildtype and Flt3ITD/+ cells. However, double MSCV-Hoxa9/MSCV-Nkx2-3 transfected
cells showed no further changes in self-renewal when compared to MSCV-Hoxa9 alone. Mean ± SEM (n=3); *p<0.05; **p<0.01; ***p<0.001; students t-test). (D) Sortng strategy for LSK/CD34+/Flt3+/CD48+ progenitor cells and overlap of diferentally expressed genes (Illumina MouseWG-6 v2 Expression BeadChip) for (i) NrasG12D/+ Vs Npm1cA/+;NrasG12D/+ and (ii) Flt3ITD/+ Vs
Npm1cA/+;Flt3ITD/+ MPPs datasets. (E) Heat map ofnormlaised Hox gene expression in puriied (i) MPP
and (ii) Lin- populatons reveal that Npm1cA/+ mutants (single or double) have similar paterns of Hox
gene expression to wildtype (normalised average expression values are used to generate heat map values). (F) Diferentally expressed genes in Npm1cA/+;Flt3ITD/+ MPPs vs wildtype controls .
Figure 3. Npm1cA and NrasG12D co-operate to drive high penetrance AML.
(A) Kaplan Meier survival curves of wildtype (n=23), Npm1cA/+ (n=34), NrasG12D/+ (n=40), Flt3ITD/+ (n=39),
Npm1cA/+;NrasG12D/+ (n=46) and Npm1cA/+;Flt3ITD+/ (n=40). Double mutant (Npm1cA/+;NrasG12D/+ and
Npm1cA/+;Flt3ITD/+) mice had a signiicantly shortened survival when compared to single mutants,
whilst Npm1cA/+;Flt3ITD had signiicantly shorter survival than Npm1cA/+;NrasG12D/+ mice. (B) Spleen and
liver weights, blood leukocyte (WCC) and platelet (Plts) counts of wildtype (n=13), Npm1cA/+ (n=17),
NrasG12D/+ (n=22), Flt3ITD/+ (n=30), Npm1cA/+;NrasG12D/+ (n=15) and Npm1cA/+;Flt3ITD/+ (n=29), Mean ±SEM,
one way ANOVA (Bonferroni adjusted) (*Vs wildtype, Vs Flt3ITD/+
, Vs NrasG12D/+). (C) Characteristc 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
histopathology from sick mice demonstrate increased incidence of AML in compound Npm1cA/ +;NrasG12D/+ and Npm1cA/+;Flt3ITD/+ mice compared to NrasG12D/+ and Flt3ITD/+ mice. Complete efacement
of splenic tssue and iniltraton of myeloid blast cells in liver tssue from Npm1cA/+; NrasG12D/+ and
Npm1cA/+; Flt3ITD+/ AMLs are presented. H&E, Haematoxylin and eosin; MPO, myeloperoxidase.
Figure 4. Leukemic progression in double mutant mice involves increased NrasG12D
or Flt3ITD
allele dosage
(A) Increase in Flt3ITD allele burden in AMLs from Npm1cA; Flt3ITD mice through loss of heterozygosity
for the locus. (i) Flt3ITD amplicon sequencing (MiSeq) of leukemic bone marrow or spleen DNA
(FN2-FN7). Tail DNA ampliied from 2-week-old Flt3+/+, Flt3ITD/+, Flt3ITD/ITD mice was used as control. (ii)
Normalised Log2 rato plots show copy neutrality of chr5 and the Flt3 locus in 7/7 Npm1cA; Flt3ITD
murine AMLs (FN-AMLs) tested. In-set: standard Flt3ITD PCR genotyping of the same FN-AML
samples; note reducton in the wildtype allele is visible. (B) (i) Summary of aCGH showing copy number gain at the Nras locus in AMLs RN6-10. (ii) Allele fractons for Nraswt vs NrasG12D show that
copy number gains in RN6-10 involved NrasG12D, and that an additonal 3 cases (RN3-5) show
copy-neutral loss-of-heterozygosity. In additon, two more RN AMLs show gains in mutant NRAS when measuring Nraswt vs NrasG12D allele fractons (aCGH was not performed on these). Results of two
Npm1cA/+ samples are also shown for comparison purposes (N6, N7). (C) Increased gene dosage of
NrasG12D correlates with increased levels of phosphorylated RAS efectors pERK1/2. FN2,3,4,6,7=
Npm1cA;Flt3ITD mice, RN1-14= Npm1cA/+;NrasG12D/+ mice.
Figure 5. Somatc mutatons in Npm1cA/+
, Npm1cA/+
; NrasG12D/+
and Npm1cA
; Flt3ITD
AMLs. (A) Exome
sequencing identies an increased number of somatc nucleotde variants (SNVs) and small indels in
Npm1cA/+, compared to Npm1cA/+; NrasG12D/+ (RN-AML) and Npm1cA; Flt3ITD (FN-AML) AML samples.
Total AMLs sequenced: Npm1cA/+ (n=12), Npm1cA/+; NrasG12D/+ (n=14) and Npm1cA; Flt3ITD (n=7),
mean/range, one way ANOVA (Bonferroni adjusted) (**=p<0.01 Vs Npm1cA/+). (B) Summary of SNVs/Indels detected in AMLs from each genotype as indicated. Those in blue are genes mutated in the TCGA AML dataset. Those in red are exact or synonymous mutatons detected in the TCGA AML dataset. (C) Co-occurance of SNVs and CNVs. Depicted are SNVs and focal copy number variatons (CNVs) which have been formally detected in the TCGA AML19 dataset or detected as common inserton sites (CIS) in our previously published Npm1cA/+ Sleeping Beauty Transposon screen 25 . Mutant allele copy gains, chromosome gains and losses depicted. For copy number variaton, colour coded boxes are based on log2 ratos (aCGH) and are not representatve of CNV size. For a complete overview of all CNV and SNV co-occurrence see Supplemental Figure S6.
535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569
A.
C.
CD1 6 /3 2 WT LK+/+ FLT3 WT LSK+/+ WT Lin-+/+ GMP CMP MEP LMPP MPP c-kit Sca-1 CD34 CD34 ST/LT (Prog./HSC) Mx1-Cre; Npm1flox-cA/+ Mx1-Cre; NrasLSL-G12D/+ Npm1flox-cA/+; NrasLSL-G12D/+; Mx1-Cre Mx1-Cre; Flt3ITD/+ Npm1cA/+; Flt3ITD/+; Mx1-Cre LSK MPP LMPP C e ll number x1 0 3 CLP C e ll number x1 0 3 LK CMP MEP GMP 0 200 400 600 800 0 50 100 150 200 250 * * ** * * ∆ ∆ ** * * ** *** † † † ** *** * * ††† ∆∆∆ * ** *** * ††† ** **** † ** ** * †E.
Npm1cA/+; NrasG12D Npm1cA/+; Flt3ITD/+ HSC MPP LMPP CLP CMP MEP GMP HSC MPP LMPP CLP CMP MEP GMPB.
cell number x 10 6 Bone Marrow Npm1cA/+ wildtype NrasG12D/+ Npm1cA/+ ; NrasG12D Flt3ITD/+ Npm1cA/+ ; Flt3ITD/+ Spleen Weight/Size *** *** ** * *** ∆∆∆ ∆∆∆ 0.0 0.2 0.4 0.6 0.8 0 20 40 60 0 20 40 60 80 s p le e n w e ig h t (g ) W C C ( 1 0 6 g/L) ** ** ** Peripheral WCCF.
C F U /3 0 ,0 0 0 B o n e M a rr o w c e lls atin g #1 atin g #2 atin g #3 0 200 400 600 ∆∆∆ ∆∆∆ * ** ♣♣♣ ♣♣♣ ∆∆∆ Npm1cA/+ wildtype NrasG12D/+ Npm1cA/+; NrasG12D Flt3ITD/+ Npm1cA/+; Flt3ITD/+D.
EPCR CD45 CD150 CD48 EPCR+ve HSC E-SLAM+/+ Percentage of viable cells 0.000 0.001 0.002 0.003 0.004 0.005 Cell number x10 3 0 5 10 15 E-SLAM HSC HSC ** * ** Npm1cA/+ wildtype NrasG12D/+ Npm1cA/+; NrasG12D Flt3ITD/+ Npm1cA/+; Flt3ITD/+ Npm1cA/+ wildtype NrasG12D/+ Npm1cA/+; NrasG12D Flt3ITD/+ Npm1cA/+; Flt3ITD/+ Npm1cA/+ wildtype NrasG12D/+ Npm1cA/+; NrasG12D Flt3ITD/+ Npm1cA/+; Flt3ITD/+ +pIpC -pIpC LK LSK46 0 0 5 2 26 1056 1 1 0 2 2 2 Hopx Hoxa5 Hoxa7 Hoxa9 Nkx2-3 Nras G12D Npm1 cA/+;Nras G12D/+ Npm1 cA/+ ;Flt3 ITD/+ Flt3 ITD/+ Lin -4926 11
A.
0.0 0.5 Npm1cA wildtype Enr ichmen t sc or e AML with NPM1 FDR q-value = 0.006 NES 2.22 0.6 0.3 0.0 Enr ichmen t sc or e Mulligan MLL Npm1cA; ; Flt3ITD wildtype FDR q-value <0.0001 NES 2.39 Npm1cA/+ Enr ichmen t sc or e Npm1cA wildtype 0.3 0.0 0.6 Leukemia with MLL FDR q-value = 0.001 NES 2.00 Npm1cA/+NrasG12D/+ Npm1cA; Flt3ITD -0.5 0.5 logFC a d j. p v a lu e -4 -2 0 2 4 10- 2 0 10- 1 5 10- 1 0 10- 5 100 HOXA genes HOXB genes NKX2-3 MEISS CD34 HOXA9 HOXA5 NKX2.3 HOXA7 HOXB5 CD34 HOXB6 HOXB3 HOXA3 Human NPM1c (n=41) Vs NPM1WT (n=28)NK-AMLB.
D.
FLT3int/hi/CD34+ CD48+CD150 -F L T 3 CD34 CD48 CD150 c K IT Sca-1 LSK LSK MPPsorted CD48+ mulipotent progenitors
normalised average expression wildtype Flt3 ITD/+ Nras G12D/+ Npm1 cA/+ ;Flt3 ITD/+ Npm1 cA/+ Npm1 cA/+ ;Nras G12D/+ Hoxa2 (i) Hoxa2 (ii) Hoxa4 Hoxa13 Hoxb5 Hoxa5 Hoxa7 Hoxa9 Nkx2-3 Hopx wildtype Flt3 ITD/+ Nras G12D/+ Npm1 cA/+ ;Flt3 ITD/+ Npm1 cA/+ Npm1 cA/+ ;Nras G12D/+ Lin -i. MPP ii. 85 21 69 Npm1cA;NrasG12D (90) NrasG12D (106)
E.
Wildtype Vs Mutant MPP 6 26 492 Flt3ITD (32) Npm1cA;Flt3ITD (518) GFP CFP CFP IRES 5’LTR Nkx2-3 3’LTR MSCV-HoxA9-GFP MSCV-Nkx2-3-CFP GFP IRES 5’LTR ψ+ Hoxa9 3’LTR ψ+expansion in liquid culture (7 days)
FACS sorting retroviral transduction
isolate lin- bone marrow re-plate
semi-solid media wildtype Flt3ITD/+
C.
i. i. ii. -5 0 5 10-10 10-5 100Also Differentially expressed in Flt3ITD/+ Chromatin/histone
Distinct to Npm1cA/+;Flt3ITD/+ MPPs
Cell Differentiation MAPK pathway JAK/STAT pathway Gldc Gdf10 Dusp6 Socs2 Cish Rasl11b Rasgrp1 Ccnd1 Jak3 Csf3r Kdm2b Chd3 Thy1 Id1
adj. p value Lmo2
Cbx7 Bcl6
Stat5a
Wildtype Vs Npm1cA/+;Flt3ITD/+ MPP
C F U /2 5 0 0 c e lls 0 20 40 60 80 wildtype Flt3ITD/+ empty vector MSCV-Nkx2-3-CFP MSCV-Nkx2-3-CFP; Hoxa9-GFP MSCV-Hoxa9-GFP ii. * *** * * * ** ** *
F.
A.
C.
Days P e rc e n t s u rv iv a l 0 250 500 750 0 20 40 60 80 100 Vs NrasG12D p<0.0001B.
AML+ AML-MPN AML-Npm1cA/+; NrasG12D/+ n=27 Npm1cA/+; Flt3ITD/+ n=11 AML+ MPN Lymph Npm1cA/+ n=12AML+ (AML with maturation) AML- (AML without matutration) MPN ( myeloproliferative neoplasm) MLL (mixed lineage leukeamia) Undiff. L. (undifferentiated lLeukeamia) Non-heam. (non-heamatological) Lymph. (lymphoid) 58% 40% 45% 100% AML+ Lymph MPN MLL NrasG12D/+ n=2213% AML+ Undiff.L Flt3ITD/+ n=12 41% Npm1cA/+ wildtype NrasG12D/+ Npm1cA/+; NrasG12D/+ Flt3ITD/+ Npm1cA/+; Flt3ITD/+ W C C ( 1 0 6/L) 0 100 300 500 s p le e n /b o d y w e ig h t r a ti o (%) 0 2 4 liv e r/ b o d y w e ig h t r a ti o (%) 0 5 10 15 Plts( 1 0 6/L) 0 500 1000 1500 2000 *** ∆ ♣ *** *** *** * ∆∆ ♣♣ *** *** ♣♣♣ *** *** * ** * ** *** Npm1cA/+ wildtype NrasG12D/+ Npm1cA/+; NrasG12D Flt3ITD/+ Npm1cA/+; Flt3ITD/+ 100µM 250µM
Npm1
cA/+;Flt3
ITD/+ Spleen Liver AML-100µM Spleen 250µM 100µM 250µMNpm1
cA/+;Nras
G12D/+Liver Spleen Liver
AML+ AML-H&E MPO Vs Flt3ITD p<0.0001 Vs Npm1cA/+ ;NrasG12D p<0.0001
A.
Npm1cA; Flt3ITD MiSeq Flt3 +/+ Flt3 ITD/+ Flt3 ITD/ITDFN2FN3FN4FN6FN 7 0.00 0.25 0.50 0.75 1.00 Flt3 ITD Flt3WT sample Flt3 ITD VAFC.
pERK1/2 RN2 (1 x NrasG12D) Control (Nras+/+) RN5 (2 x NrasG12D) RN6 (4 x NrasG12D) ERK1/2 x80 x40 x80 x40 x80 x40 x80 x40 x80 x40 x80 x40 x80 x40 x80 x40 x40 x40 Npm1cA/+; NrasG12D/+i.
i fl i lo g 2 r a ti o 5:1.4 0x10 8 5:1.4 4x10 8 5:1.4 7x10 8 5:1.4 9x10 8 -0.75 -0.50 -0.25 0.00 0.25 0.50 0.75 chromosome 5 Flt3 copy neutralii.
aCGH NrasG12DB.
Nras G12D VAF N6 N7RN1RN2RN3RN4RN5RN6RN7RN8RN9 RN10 RN11RN12RN13RN14 0.00 0.25 0.50 0.75 1.00 + 3 +1 + 0 .5 +3 +3 c .n . c .n . c .n . c .n . c .n . c .n . c .n . no aCGH sample MiSeq N ra s copy number N6 N7 RN 1 RN2RN3RN4RN5RN6RN 7 RN 8 RN9RN10 0 1 2 3 4 5 6 +3 +2 +1 diploid aCGH Npm1cA/+; NrasG12D/+ Nraswt samplei.
ii.
A.
B.
0 5 10 15 ** ** Av e ra a g e # ofSNVs and Indels per AML
Npm1 cA/+ N pm1 cA/+ ; Flt3 ITD/+ N pm1 cA/+ ; Nras G12D/+ Mtmr2 p.R621H Psma7 p.I188S Vmn1r113 p.S142C Cdv3 p.N272K Chi3l4 p.N279T Speer4b p.H49L Gm4297 p.Y199S Nlrp4d p.I299L Igsf1 p.V411M Klra4 p.V190L Kif3c p.R504L Col4a3 p.P185L Spry2 p.V61E Kdm5b p.Q962* Pik3r1 p.D560N Olfr201 p.S87C Mrps18c p.K14R Gm6316 p.K149R Fsip2 p.T6433M Zfp111 Exon2+2 Atm p.E1789ED>D 1810030J14Rik Gm5422 p.M266V Prl8a6 p.T105M Cass4 p.A454V Arhgap20 p.R26L Cep72 p.G642S Bcl9l p.G312D Rasa3 Exon14+2 Ptprs p.T485P Dsp p.T1276M Olfr784 p.K54M Fdft1 p.R185C Myh3 p.R1251G Krt34 p.E327K Mup14 p.R57K Pkd1l1 p.S337R Ptpn11 p.S502L Nf1 p.R683* 1300010F03Rik p.G1730R Zbtb40 p.G332E
Stag1 p.V589I Fam35a p.F204fs*490
Ccdc19 p.A62T Smarcc1 p.E757H AC127247.1 p.R90H Zfp93 p.K320R Rabac1 p.Q52R Cd207 p.D138V Atp13a4 p.T942M Kl p.V190G Kif3c p.R504L Gldn p.D329G Irgq p.V325I Vwa2 p.T622M Epb4.2 p.A268V Ptpdc1 p.R280fs*24 Ptpn12 p.N58fs*3 Xaf1 p.R134fs*17 4931417E11Rik p.A211T Npm1cA/+ Eif4a3 p.D22G Ldhb p.R170Q Smek1 p.V8G Cdc7 p.N180K Ccni p.W51* Tmem92 p.P91S Grin2a p.R518C
Myst3 p.V944F Olfr156 p.V44I
Hmha1 p.R706H Dnahc7a p.I1854V Ccdc53 p.S104L Mapk8ip3 p.R42H Trpc1 p.R314Q Col8a2 p.P469L Hebp1 p.V86L Gm6723 p.G41A Net1 p.G514S Upf2 p.E1041K Fcrl1 p.I128N Stab1 p.V2370I Ube2c p.C114S Fsip2 p.H777Y Cd300lb p.W149 Pi4k2a p.V148I Adamts2 p.R1017* Gm7030 p.H108N Mex3d p.V224E Vmn1r169 p.R267W Akap9 p.G3302A Fibcd1 p.P143H 6430527G18Rik p.N551K Chd5 p.R1298L Serping1 p.N83K Pde4dip p.R478Q Mcph1 p.E494K Zdhhc4 p.R313W Cpsf1 p.V4G Espn p.T134P Nhlrc1 p.K201fs*33 Tmem164 p.G26fs*43 Serpinb6e p.S243Ins(KS Ptpru p.L72V Isl2 p.V234G Spata2l p.P369S Nlrp4d p.L727P Eefsec p.V6G Clstn1 p.T741PPcdh8 p.T771P Ccdc30 p.D301G F13a1 p.G411A AW554918 p.D430N Col24a1 p.L1453P Net1 p.G514S Gucy2c p.V1006I Nf1 p.W1260* Ankrd35 p.R745W Hdac2 p.K10N Bsn p.G3337R Nup214 p.G2017fs*134 Fat3 p.R3621H 5430419D17Rikp.T5P Mga p.V1800G Pcdhb16 p.T611M Flt3 p.D842G Slit2 p.C502R Ptpn11 p.R498W Slc2a7 p.R177K Nras p.Q61H Dock8 p.T29fs*29 Ikzf1 p.P244fs*4 Idh1 p.R132Q Ikzf4 p.P426L Npm1cA/+; Flt3ITD/+ Npm1cA/+; NrasG12D/+ Wac p.V541A Cbl p.S374F Ptpn11 p.E69K Npm1cA/+ 1 2 3 4 5 6 7 8 1 234 5 6 7 8 9 1 2 3 4 5 6 7 Npm1cA/+; NrasG12D Npm1cA/+; Flt3ITD/+ chromosome 2 chromosome 6 chromosome 7 chromosome 10 chromosome 15 chromosome 17 NrasG12D Flt3ITD Nras Ptpn11 Cbl Idh1 Ikzf4 Ghr Ezh2 Wac Sfi1 Sh2b3 Crocc
mutant allele duplication/amplification whole or partial chromosomal gain single copy of mutant allele SNV/Indel 1.0 -1.0 0.1 -0.1 Log2 Ratio Dock8 aCGH Gucy2 Mpl Myo5b Fat3 Pkd1l2 Wac