Comparative clonal analysis of reconstitution kinetics after transplantation of hematopoietic stem cells gene marked with a lentiviral SIN or a γ-retroviral LTR vector

Experimental Hematology 2013;41:28–38

Comparative clonal analysis of reconstitution kinetics after transplantation of hematopoietic stem cells gene marked with a lentiviral SIN or a g-retroviral LTR vector Kerstin Cornilsa, Cynthia C. Bartholomaeb, Lars Thieleckec, Claudia Langea, Anne Arensb, Ingmar Glauchec, Ulrike Mocka, Kristoffer Rieckena, Sebastian Gerdesc, Christof von Kalleb, Manfred Schmidtb, Ingo Roederc, and Boris Fehsea a

Research Department Cell and Gene Therapy, Clinic for Stem Cell Transplantation, University Medical Centre Hamburg-Eppendorf, Hamburg, Germany; bDepartment of Translational Oncology, National Centre for Tumour Diseases and German Cancer Research Centre, Heidelberg, Germany; cInstitute for Medical Informatics and Biometry, Faculty of Medicine Carl Gustav Carus, Dresden University of Technology, Dresden, Germany (Received 17 January 2012; revised 28 August 2012; accepted 10 September 2012)

Retroviral gene marking has been used successfully in preclinical and clinical transplantation settings. Highly sensitive techniques for vector insertion-site determination, such as linear amplification–mediated polymerase chain reaction (LAM-PCR) in conjunction with nextgeneration sequencing, have been introduced to assess the composition of gene-marked hematopoiesis at a single-cell level. Here we used these novel techniques for directly comparing clonal reconstitution kinetics in mice transplanted with bone-marrow–derived stem cells genetically marked with either a standard, spleen focus–forming virus long terminal repeat (LTR)–driven g-retroviral, or a lentiviral self-inactivating vector containing an identical but internal spleen focus–forming virus–derived enhancer/promoter. We observed that the use of the lentiviral self-inactivating vector for gene marking was associated with a broader repertoire of differently marked hematopoietic clones. More importantly, we found a significantly higher probability of insertions in growth-promoting, clonal-dominance–associated genes in the spleen focus–forming virus LTR–driven g-retroviral vector at later time points of analysis. Based on our data, we suggest that the combined use of LAM-PCR and nextgeneration sequencing represents a potent tool for the analysis of clonal reconstitution kinetics in the context of gene marking with integrated vectors. At the same time, our findings prove that the use of multiple restriction enzymes for LAM-PCR is indispensable to detect most or ideally all individual stem cell clones contributing to hematopoiesis. We have also found that techniques such as quantitative PCR can be helpful to retrospectively analyze reconstitution kinetics for individual hematopoietic stem cell clones. Finally, our results confirm the notion that marking with lentiviral self-inactivating vectors is associated with a lower risk of genotoxicity as compared with g-retroviral LTR vectors. Ó 2013 ISEH - Society for Hematology and Stem Cells. Published by Elsevier Inc.

Gene marking with integrating, retroviral vectors has been very useful to follow-up the in vivo fate of cells. In fact, important insights in stem cell biology, but also the potential impact of the gene-marking vectors, have been Offprint requests to: Boris Fehse, Ph.D., Research Department Cell and Gene Therapy, Clinic for Stem Cell Transplantation, University Medical Centre Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany; E-mail: [email protected] Supplementary data related to this article can be found online at http:// dx.doi.org/10.1016/j.exphem.2012.09.003.

obtained in various animal models [1–5]. Gene-marking approaches in clinical cell therapy settings contribute significantly to our understanding of highly relevant phenomena, such as relapse induction after stem cell transplantation or homing potential of ex vivo expanded immune cells toward tumors [6,7]. In clinical gene therapy, cure of the disease was accompanied or followed by severe adverse events in some patients [8–12]. Remarkably, malignant transformation was related to the up-regulation of cellular proto-oncogenes by proviral regulatory elements. These findings emphasize the need for uncovering the transduced

0301-472X/$ - see front matter. Copyright Ó 2013 ISEH - Society for Hematology and Stem Cells. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.exphem.2012.09.003

K. Cornils et al./ Experimental Hematology 2013;41:28–38

cell pool and dissecting the contribution of individual clones to hematopoietic repopulation. Recently, large-scale clonal analyses of gene-marked cells have been revolutionized based on novel technique combinations, specifically ligation-mediated or linearamplification-mediated (LM/LAM-) polymerase chain reaction (PCR) and next-generation sequencing (NGS) [13–17]. Using LM-PCR followed by pyrosequencing, the Baum laboratory investigated the impact of target cells (hematopoietic stem vs progenitor cells), vector types (g-retroviral vs lentiviral vectors) and ex vivo culture conditions on the probability of clonal dominance [18,19]. However, LM-PCR has been shown to preferentially retrieve insertion sites present in dominant clones [3,20,21]. In contrast, LAM-PCR facilitates retrieval of insertion sites present at low frequencies in the cell pool, although constraints exist due to the use of restriction enzymes and the PCR steps favoring shorter fragments [22–25]. We have previously introduced murine stem cell transplantation and marking models for in vivo analysis of retroviral gene transfer vectors [26,27]. These models have been shown to be predictive for unwanted side effects of retroviral gene transfer, such as vector-induced clonal dominance or leukemia [2,3,20]. Here we combined the established model with techniques mentioned for highresolution insertion site analysis to directly compare reconstitution after transplantation of hematopoietic stem (HSC)

A

Lenti-SF

SIN-LTR

SF91

SFFV-LTR

29

and progenitor cells (HPC) transduced by either a standard g-retroviral spleen focus–forming virus (SFFV) vector (SF91) or a lentiviral self-inactivating (SIN) vector with an internal SF91 U3 promoter/enhancer (termed LentiSF). In order to analyze engraftment kinetics, we obtained blood samples from transplanted animals at regular intervals (approximately every 4 weeks). On these samples, we performed LAM-PCR, followed by 454 sequencing to uncover and quantify clonal contribution of gene-marked cells. We show that by using these techniques, clone-specific reconstitution kinetics by gene-marked stem and progenitor cells can be analyzed in detail. At the same time, our data indicate that application of additional techniques, such as quantitative PCR, might be necessary for more detailed analyses of individual clonal contributions. Finally, comparative analysis of the two vector types in our sensitive mouse model corroborates an improved safety profile of lentiviral SIN vectors.

Materials and methods Vector design and production The SF91 vector was described earlier [2]. The Lenti-SF– enhanced green fluorescent protein (eGFP)-vector is an SIN lentiviral vector that contains an eGFP gene under the control of an internal spleen focus–forming virus (SFFV) promoter (Fig. 1A). It essentially resembles the structure of LeGO-G2 [28] with two

SFFV

eGFP

eGFP

wPRE

wPRE

SIN-LTR SFFV-LTR

B

7 months

, CD45.2

, CD45.1 BloodSamples

Figure 1. Vector design and experimental set-up. (A) The SF91-eGFP vector was described previously [26]. The Lenti-SF-eGFP vector was constructed by cloning the SFFV-promoter element from the SF91 into the LeGO-G2 vector [28]. (B) Viral supernatants of the two vectors were used to transduce lineagedepleted bone marrow cells from male CD45.2 mice. Transduced cells were transplanted into lethally irradiated female CD45.1 recipients (n 5 10). Animals were observed for 7 months. During that observation time blood samples were taken every 4 to 5 weeks.

30

K. Cornils et al./ Experimental Hematology 2013;41:28–38

additional modifications: the SIN-long terminal repeat (LTR) contains the large deletion described by Zufferey et al. [29], and the wPRE element was safety optimized [30]. Cell-free supernatants containing viral particles pseudotyped with mouse Moloney ecotropic envelope protein (env) were obtained by transient transfection of 293T cells [27,31]. Vector titers were determined on SC-1 cells as described previously [28]. Bone marrow preparation, stimulation, and transduction Bone marrow cells were harvested from femora and tibiae of 8- to 10-week-old male C57Bl/6 mice (CD45.2). Lineage depletion was performed according to manufacturer’s instructions (Miltenyi Biotec, Bergisch-Gladbach, Germany). The lineage-negative fraction of cells was prestimulated in StemSpan Medium (StemCell Technologies, Grenoble, France) containing 100 ng/mL of the following cytokines: murine stem cell factor, human thrombopoietin, human fms-like tyrosine kinase receptor-3 ligand, and human interleukin-11 (Immunotools, Friesoythe, Germany). After 2 days of stimulation, cells were transduced with either the SF91 or the Lenti-SF vector (both ecotropic pseudotypes) on Retronectin- (TaKaRa, Otsu, Shiga, Japan) coated plates at a multiplicity of infection (MOI) of 3 (Fig. 1B). Transduction efficiencies were estimated via fluorescence-activated cell sorting (FACS) analysis. Transplantation and blood sampling Lethally irradiated (10 Gy) C57Bl/6 female mice (CD45.1) were transplanted with 5
105 cells via tail vein injection (n 5 10 per group). After transplantation and during follow-up, three mice per group died due to unknown reasons, most probably engraftment failures. During the observation time of 7 months, serial peripheral blood samples were taken via retro-orbital bleeding approximately every 4 weeks starting 6 weeks after transplantation (Fig. 1B). A small amount of blood was used to monitor the reconstitution by analyzing the eGFP-expression via FACS. The remaining blood cells were used for DNA extraction (Qiagen, Hilden, Germany). Four animals per group were randomly chosen for LAM-PCR and 454 sequencing. After 7 months, all animals were sacrificed and bone marrow, blood, spleen, and thymus were collected for further analysis. LAM-PCR and 454 sequencing One hundred nanograms DNA were used for LAM-PCR as described by Schmidt et al. [22]. Two restriction enzymes (SF91: Tsp509I and HinPI; Lenti-SF: Tsp509I and HpyCH4IV; all NEB, Ipswich, MA, USA) were used for each vector to increase the recovery of integration sites in independent LAM reactions [14]. According to Gabriel et al. [24], both restrictionenzyme combinations facilitate very similar genome accessibility (Lenti-SF: 56%, SF91: 54%). LAM-PCR products were used for 454 pyrosequencing after introduction of barcoded primers, purification of the evolved product, and quantification [32]. LAM products independently generated for each blood sample (using the two different restriction enzymes) were sequenced. Sequencing was performed in the German Cancer Research Centre core facilities. To obtain a representative picture of the clonal composition of the blood at each time point, we aimed at assessing approximately 1000 sequences per sample. The actual range of processed sequences eligible for clonal analyses was 53 to 2203 reads at different time points.

Sequence analysis and clonal distribution Sorting of the sequences according to their sample-specific barcode and trimming to remove LTR and linker-sequences was done using a proprietarily developed in-house Perl script by the Department of Translational Oncology, National Centre for Tumour Diseases and German Cancer Research Centre (Heidelberg, German) [32]. Quality control included checking the obtained sequences for the presence of the last PCR primer (LTR3 for LentiSF and A3 for the SF91 [22]) and at least 5 bp of the viral LTR sequence. For automatic analysis of the presorted sequences, a specific RScript was developed using sophisticated clustering and mapping software. In the first step, the single reads were mapped to the mouse genome (July 2007, NCBI37, mm9) via Segemehl Vers. 0093 [33]. Sequences with !85% of similarity to the mouse genome were excluded from further analysis. Those sequences were mostly showing the internal control of the LAM-PCR, other vector-derived sequences and some other sequences, which can be attributed to an inherent technical issue of NGS termed collisions [34]. Also, single reads were excluded from analysis. The remaining sequences were clustered with UClust (Version 1.1.579, [35]) to identify clusters of identical sequences. Assignment of the corresponding genomic integration sites to the specific clusters was done by calculating the cluster-specific consensus sequence using clustal 2.0.12 [36] and submitting these sequences to a virus integration-site analysis tool developed at the Medical School Hannover (http://eh.mh-hannover.de/isa). A window of 6250 kb around the identified integration site was analyzed for the closest gene. Identified genes were checked for their presence in the retroviral tagged cancer gene database (RTCGD) [37]. LM-PCR and quantitative real-time PCR LM-PCR using 200 ng DNA as a template was described previously [38]. The same restriction enzymes were used as for LAM-PCR (two per vector, see LAM-PCR and 454 sequencing). Two potentially important integration sites identified by LMPCR on spleen DNA were chosen to assess clonal contribution of the corresponding clones in all respective blood samples via quantitative real-time PCR (qPCR)done identified in animal I.4.5 (Mcts1), the other one in animal I.4.7 (Trim32). The LTRgenome junction of the corresponding clone was PCR amplified (Taq-Polymerase, Fermentas, St. Leon-Rot, Germany) using integration-specific primers (LTR-specific forward-primer: 50 CCCGTCTGTTGTGTGACTCT and the integration-specific reverse-primer; Mcts1-RV: 50 -CTCTGTGTAGCCCTGGCTGT and Trim32-RV: 50 -GCAGATCTCTTTAGCTCCTTGG). The obtained fragments (169 bp for Mcts1, 113 bp for Trim32) were cloned into the pCR 2.1-TOPO vector (TOPO cloning kit; Invitrogen, Carlsbad, CA, USA) to generate plasmid standard curves (108 to 103 copies/mL) as described by Bozorgmehr et al. [39]. In order to determine the relative contribution of the specific clone at a given time point, we performed a control PCR detecting a 109-bp fragment derived from an intron of the Erythropoietin receptor (EpoR) gene in parallel. We used the following primers: EpoR-FW: 50 -GCAGGCGGGGTCGCTACTC and EpoR-RV: 50 CGC CTGTGCAGATCCGATAA. Before SYBR-Green-based qPCR (QuantiTect SYBR Green; Qiagen), the DNA of the blood samples was amplified via Whole Genome Amplification (Repli-g; Qiagen). The total amount of copies of the clone was assessed by absolute quantification from

K. Cornils et al./ Experimental Hematology 2013;41:28–38

the respective plasmid standard curve. Contributions of the Mcts1clone in animal I.4.5 and the Trim32-clone in animal I.4.7 were calculated by division of the absolute copy numbers of the clone by the absolute copy numbers of the EpoR gene.

31

Despite applying equalized MOIs, transduction rates were quite different for the two vectors. While approximately 74% of the lineage-depleted bone marrow cells were transduced with the SF91 vector, the corresponding rate was 18% for Lenti-SF (Fig. 2A). In both cultures, the proportion of earlier, c-kit–positive cells was almost equal (34% vs 31%). Accordingly, the percentage of eGFPpositive c-kit cells was also more than four times higher in the SF91 group (26% vs 6%; Fig. 2A). Remarkably, copy numbers of proviral vector DNA estimated by qPCR 24 hours after transduction, i.e., before integration, were essentially equal for both vector types (SF91: 2.34; LentiSF: 2.45). These data indicate that the applied MOIs were well adjusted, and facilitated entry of identical amounts of viral particles. However, mean numbers of integrated

Results Efficient transduction of murine HSC/HPC with both vectors In order to ensure low vector copy numbers per transduced cell (i.e., 1–2), we aimed at an intermediate gene transfer efficiency of approximately 30% to 50% [40]. To this aim, we used equal MOIs of 3 for both vectors. MOIs were calculated based on vector titers determined on murine SC-1 cells.

A Lenti-SF

104

6

57

25

26

18

8

102

48

1011

1022

1033

1044

100

101

101 100

100

103

12

102

103

GFP

104

SF91

100

1011

1022

1033

1044

c-kit

B 100

% eGFP

80 60

SF91 Lenti-SF

40 20 0 6

10

15

23

28

weeks Figure 2. FACS analysis of freshly transduced and peripheral blood cells to detect eGFP expression. (A) Analysis of transduced cells by flow cytometry for expression of eGFP and the stem cell marker c-kit (phycoerythrin-coupled antibody) 48 hours after transduction. As evident, transduction efficiency in the SF91-transduced cells was approximately 4 times higher in comparison to Lenti-SF–transduced cells. (B) Percentage of eGFP-expressing cells in wholeblood samples taken at indicated time points post-transplantation. Relative numbers of eGFP-expressing cells remained stable over the observation time in both experimental groups (n54 randomly chosen animals per group).

32

K. Cornils et al./ Experimental Hematology 2013;41:28–38

vector copies detected at final analysis in vivo in spleen cells were, in full agreement with the FACS data fourfold higher for SF91: 1.06 (n 5 4) than for Lenti-SF: 0.23 (n 5 4). Also, similar donor chimerism as assessed by flow cytometry was observed for the two vector groups (LentiSF: 65.5%; SF91: 55.3%). While the reason for the relatively big difference in transduction rates could not be resolved unambiguously, sufficient numbers of cells were transduced for both vectors to allow for later follow-up of vector-positive cells in vivo. This was confirmed by determining numbers of eGFPpositive cells using FACS of whole-blood samples at each data point during the entire observation time (Fig. 2B), as well as by a detailed analysis of eGFP contents in different blood cell lineages at final analysis (Supplementary Table E1; online only, available at www.exphem.org). Numbers of clones contributing to hematopoiesis over time are relative constant for both vectors In the next step, we used LAM-PCR followed by NGS to determine relative clone sizes at each given time point of analysis. To do so, all obtained sequences underwent processing and quality control. Only those sequences that fulfilled initial quality criteria and could be mapped with a similarity of at least 85% to the mouse genome were included in further analysis. Together w40,000 sequences were obtained and O30,000 sequences passed all quality criteria (Table 1). Unfortunately, no blood samples could be obtained for LAM-PCR at week 28 for mouse I.4.5, Table 1. Numbers of obtained sequence reads by 454 sequencing after trimming and additional quality control After trimming Mouse

W6

W10

W15

W19

W23

W28

Total

I.1.2 I.1.4 I.1.5 I.1.6 I.4.2 I.4.4 I.4.5 I.4.7 Total

103 1092 1334 504 2549 590 755 973

126 1405 886 1062 131 772 1092 1040

125 2086 965 928 158 754 1236 823

170 150 530 1321 131 150 754 628

1126 1155 1223 956 2704 258 994 1355

1413 1212 1394 435 38 37 0a 824

3063 7100 6332 5206 5711 2561 4831 5643 40,447

1271 1127 1354 428 29 0 0a 126

2431 5698 5862 4553 4457 1774 3228 2930 30,933

After additional quality control I.1.2 I.1.4 I.1.5 I.1.6 I.4.2 I.4.4 I.4.5 I.4.7 Total a

53 675 1171 452 2203 408 635 455

67 845 823 833 66 544 757 867

84 1977 835 632 78 528 983 448

No blood sample available.

59 116 521 1306 66 62 528 353

897 958 1158 902 2015 232 325 681

and LAM-PCR was not efficient for mice I.4.2 and 1.4.4. Therefore, cluster data could not be compared at that time point between the two groups. Using the clustering algorithm described in Material and Methods, identical and highly similar sequences were clustered. Based on this, numbers of individual clusters at each time point were assessed for both vectors for each individual animal. Based on the assumption that a given cluster represents a single clone with a defined insertion site, this data was used to assess the kinetics of clone numbers contributing to hematopoiesis (Fig. 3). To allow for direct comparison between the two vector groups, numbers of clusters were related to the absolute number of genemarked cells. Results shown in Figure 3 reflect the calculated mean numbers of different clones per 100,000 transplanted, gene-modified cells in the peripheral blood of each of four mice per vector group at the indicated time points after transplantation. The obtained data indicate that, despite large inter-animal variation in both vector groups, the numbers of clones contributing to hematopoiesis were relatively stable. Patterns of clone size changes during hematopoietic recovery with different marking vectors In the next step, the relative contribution of each clone to (the gene-marked proportion of) hematopoiesis was determined based on the sequence numbers (read counts) for a given cluster in relation to the whole number of relevant sequences obtained for that time point. These data were used to create kinetic plots reflecting hematopoietic reconstitution of each individual animal by gene-marked cells. For plainness, relative data obtained for all animals (n 5 4) of each of the two vector groups were merged into one graph (Fig. 4). As illustrated, contribution of individual clones to hematopoiesis followed wave-like kinetics for essentially all clones marked with Lenti-SF (Fig. 4B). On the contrary, single clones in the SF91 group apparently showed a tendency for continuous expansion and remained dominant over time (Fig. 4A, upper clone). Long-term reconstituting clones marked with SF91 show a tendency toward over-representation of insertions in the vicinity of proto-oncogenes Next we used the sequence data obtained by NGS to identify and compare the distribution of vector-insertion sites in long-term reconstituting clones marked with either SF91 or Lenti-SF. We focused on insertions into genes previously included in the RTCGD (Supplementary Table E2; online only, available at www.exphem.org) [36]. Four mice were available for analysis per group. For comparison of the two vector groups, we determined for each time point the number of clusters representing insertions in the proximity of RTCGD genes relative to the absolute number of clusters (Fig. 5A). A relatively large window of 6250 kb around the integration site were used for gene analysis. Notably,

K. Cornils et al./ Experimental Hematology 2013;41:28–38

33

25.0

clusters / 100.000 cells

Lenti-SF

SF91

20.0

15.0

10.0

5.0

0.0 6

10

15

19

23

weeks Figure 3. Mean numbers of clusters per 100,000 transplanted, transduced cells in the peripheral blood of individual mice at different time points. Numbers of clusters were determined after processing the obtained sequences with a cluster algorithm (compare Materials and Methods). Numbers of clusters were calculated for individual animals per 100,000 transplanted, eGFP-positive transduced cells based on the initial transduction rate of 0.74 for SF91 and 0.18 for Lenti-SF. Data represents mean values for four animals per group.

already at early time points after transplantation (weeks 6 and 10), the relative contribution of RTCGD clusters to the total number of clusters was lower for Lenti-SF (e.g., at week 6 by a factor of 2.4), although differences were not significant (p 5 0.28; Fisher’s exact test). These data most probably reflect the different integration preferences of the two vector types. Strikingly, the relative contribution of RTCGD clusters remained at high levels of about 20% over the whole observation period in the SF91 group, while it decreased to approximately 2% in the Lenti-SF group at

the final data point. At final analysis, the difference between the two groups was highly significant (p 5 0.003; Fisher’s exact test). Also, the overall relative representation of RTCGD hits was significantly higher in the SF91 as compared with the Lentis-SF group (p ! 0.05; Fisher’s exact test). Finally, we analyzed the relative proportion of sequence counts for RTCGD insertions in comparison to the total read counts at any given time point (Fig. 5B). Again, the relative contribution of clones representing RTCGD hits was significantly larger in the SF91 group as

Figure 4. Clonal composition of hematopoiesis over time. Kinetics of relative contribution of individual sequence clusters (also referred to as clones) to hematopoietic reconstitution after transplantation are shown for the two vector groups. Data shown reflect a summarized analysis for each of the four animals per group. Relative clone sizes were adjusted to the total sequences obtained for each individual mouse at each data point.

34

K. Cornils et al./ Experimental Hematology 2013;41:28–38

A Relative Proportion of RTCGD-cluster (%)

30.0

Lenti-SF

p=0.011

SF91

25.0

p=0.003 20.0

n.s.

n.s.

n.s.

15.0

10.0

5.0

0.0 6

10

15

19

23

weeks

B

40

Mean Fraction of RTCGD-reads (%)

Lenti-SF

p=0.043

SF91

n.s.

35

n.s.

30

n.s. 25

n.s. 20 15 10 5 0 6

10

15

19

23

weeks Figure 5. Mean numbers of clones/clusters containing insertion sites in the proximity of retrovirally tagged cancer genes and their abundance in the blood of mice over time. Cluster sequences were blasted against the mouse genome; genes in a window of 6250 kb were identified and analyzed for their presence in the RTCGD [37]. (A) Mean numbers of RTCGD clusters in relation to total cluster numbers are depicted for the four animals per vector group and the different time points. (B) Read counts for the RTCGD-clusters [as in (A)] were set in relation to the total reads for the four individual mice at all time point; depicted are the means for both groups; whiskers indicate the standard error; p values indicate significant differences between the two groups, n.s. 5 not significant.

compared with the Lenti-SF group at the final point of analysis (Welch two-sample t test). Remarkably, at later time points, RTCGD hits apparently capture a higher share of

hematopoiesis per cluster (based on the relation of read counts to cluster numbers) in the SF91, but not the LentiSF group. For example, the relative quota of RTCGD

K. Cornils et al./ Experimental Hematology 2013;41:28–38

clusters for SF91 at week 23 is approximately 18%, and the respective reads for these clusters total up to 28%. In contrast, for Lenti-SF group, the relative proportions of RTCGD hits and counts are essentially identicaldboth approximately 2% (Fig. 5). In line with this, at week 23 RTCGD hits were present in all four animals of the SF91 group (representing up to 42% of total read counts in single individuals), but only in one of four Lenti-SF mice (3% of read counts). These data again support the notion that clonal dominance due to insertional mutagenesis is more likely to develop after transduction with SF91 than with Lenti-SF. LAM-PCR þ NGS provide a representative clonal analysis At final analysis, we performed LM-PCR [38] on spleen DNA samples. This enabled us to reveal single insertion sites (e.g., insertion I.4.5, Fig. 6B) that had escaped detection in peripheral blood (not shown). As reported previously, a restriction fragment length bias becomes relevant when using LM- and/or LAM-PCR [23,24,38]. Most probably due to the latter, only the 276-bp Tsp509I- but not the 748-bp HpyCH4IV fragment of the respective insertion in mouse I.4.7 (Fig. 6A) was detectable by LAM-PCR at different time points. Accordingly, insertion in mouse I.4.5 was detected by LAM-PCR/NGS at only one time point due to the large fragment sizes generated by both HpyCH4IV and Tsp509I (Fig. 6B).

In line, quantification of LAM products does not always correlate with results from qPCR using insertion-specific primers. In fact, we observed a nice correlation for the Trim32-clone, which is obviously based on the efficient amplification of the short Tsp509I-fragment (Fig. 6A). However, insertion into the Mcts1-locus remained almost unnoticed by LAM-PCR at each time point analyzed, which strongly contrasted qPCR results. Together these in vivo data confirm the prediction that even the use of two different restriction enzymes for LAM-PCR may not strictly ensure a complete figure of gene-marked clones [23,24]. Alternative techniques, such as nonrestrictive LAM-PCR, have been developed to solve this problem. In addition, insertion-site–specific qPCR techniques can be helpful for validating quantification data.

Discussion In this work, we used an established bone marrow transplantation assay to directly compare clonal reconstitution kinetics after transplantation of hematopoietic stem and progenitor cells marked with either a g-retroviral LTR vector or a lentiviral SIN vector with an internal promoter. Hematopoietic reconstitution was assessed over time based on blood samples taken every 4 weeks (starting 6 weeks after transplantation). Sensitive detection methods, namely LAM-PCR followed by 454 NGS were applied in order to ensure the highest possible sensitivity of clone detection.

B

mouse I.4.7 1.2

1.2

0.20

0.20

454-reads

1

0.8

0.8

0.6

0.6

0.4

0.4

0.2

0.2

copy ratio insertion/hk

mean qPCR

read counts

copy ratio insertion/hk

mean qPCR 1

mouse I.4.5

454-reads 0.15

0.15

0.10

0.10

0.05

0.05

0.00 0 6

10

15

19

23

28

Tsp509I HpyCH4IV 276, 314 bp 748 bp

10

15

19

23

weeks

weeks

3‘ LTR

0.00 6

0

read counts

A

35

3‘ LTR

Tsp509I 514, 870 bp

HpyCH4IV 2636 bp

Figure 6. Comparative quantitative analysis of LAM-PCR amplicon sequence counts and clone-specific qPCRs. We directly compared quantitative analysis of clonal contribution as determined by LAM-PCR and NGS (relative read counts/cluster numbers) and integration-specific qPCR (copy ratio). Two different examples of outcomes are shown. (A) Integration-specific quantitative real-time PCR for an integration found in mouse I.4.7 was established. Comparison of qPCR data with quantitative cluster analysis revealed a very good correlation. In line, analyses of the restriction sites of the used enzymes near the insertion locus showed a good accessibility at least for Tsp509I. (B) On the contrary, only weak correlation of integration-specific qPCR and quantitative cluster analysis based on 454 sequencing was found for an integration site from mouse I.4.5. As indicated, restriction sites for both enzymes used for LAM-PCR were relatively distant from the insertion, which strongly impairs detection by LAM-PCR [24].

36

K. Cornils et al./ Experimental Hematology 2013;41:28–38

These methods were previously shown to facilitate efficient detection of vector insertion sites, particularly in clinical samples from gene-therapy studies [14,15,41]. Importantly, the following conditions have to be met in order to obtain a representative picture of clonal blood composition using the approach described here: the amount of sampled blood needs to be sufficient for analysis but low enough to avoid a significant impact on hematopoiesis; the LAM-PCR product should reflect clonal composition, which requires sufficient DNA amounts and quality; sequencing should be deep enough to catch low-abundant clones. As also noted in this work, due to technical limitations it is sometimes difficult to meet all criteria, e.g., obtain sufficient blood. Therefore, single data points had to be excluded from analyses or have to be interpreted with caution. Analyzing hematopoietic reconstitution on a singleclone level provides important insights into the regenerative potential of blood stem and progenitor cells. In addition, assessing growth kinetics of single clones over long periods of time can facilitate our understanding of malignant transformation and progression [42]. Our data confirm earlier studies indicating gradual and differential involvement of single HSC/HPC during hematopoietic reconstitution [43]. Indeed, different kinetics could be seen for individual hematopoietic clones: while some could be detected at all time points analyzed, other clones were present only early or only late after transplantation. Also, several clones presented an oscillating behavior. Finally, for some clones we observed a steady expansion over time. Notably, clonalexpansion was much more pronounced in the SF91 group. With regard to the question of clonal composition of hematopoiesis, it also needs to be taken into account that a large proportion of transplanted HSC/HPC were not gene marked. Obviously, to assess the actual number of clones contributing to hematopoiesis at any time point the numbers as defined in our work need to be extrapolated. Also, in the given experimental setting, we had no possibility of discriminating between short-living cells, such as granulocytes, and long-living cells, such as memory lymphocytes, during the follow-up study [44]. Consequently, in the given study the clonal repertoire measured at each time point represents a mixture of different cell types, not all of which mirror active HSC/HPC. In fact, contribution of marked cells to all blood lineages was confirmed at final analysis, admittedly with great variances among individual animals (e.g., from 1.4% to 73.6% for Gr1–positive cells; Supplementary Table E1; online only, available at www.exphem.org). At the same time, our marking data suggest that a significant proportion of gene-marked cells analyzed at any given time point represented short-lived cells that reflected the activity of HSC/ HPC. This is in good agreement with previous work in the given model indicating particularly efficient retroviral gene marking for the myeloid compartment [3,20,26,27].

Our study also has provided important insights into the impact of vector design on clonal contribution and evolution during reconstitution. Two types of vectors were used for HSC/HPC marking in a parallel experimentda standard g-retroviral LTR vector and a lentiviral SIN vector. For direct comparability, we used identical SFFVderived promoter sequences in both vectors. In a previous study, Kustikova et al. [18] also compared the impact of these two vector types on clonal blood composition, although not in a parallel experimental setting and based on LM-PCR. LAM-PCR followed by NGS as used in our study has been shown to have a much higher resolution power for clones represented at low frequencies [22]. Using an identical transduction protocol for both vectors, we were able to ensure sufficient gene-transfer rates into HSC and HPC. However, despite application of identical MOIs (as based on titration on SC-1 fibroblasts), initial transduction rates were significantly different. Our data indicate that a restriction for the HIV-based vectors occurred at the postentry/integration level in the murine bone marrow cells, which had not been observed in murine SC-1 cells used for titration. Notably, three- to fourfold variations in integration efficiencies were previously found for VSVpseudotyped lentiviral constructs in different mouse cell lines [45]. Those differences can most probably be ascribed to some cell-type–specific postentry/preintegration blocks. Indeed, specific blocks for HIV integration were found in murine T cells, for example [46]. We could also not exclude that lentiviral vectors pseudotyped with the mouse envprotein are less infectious than g-retroviral vectors with the same env protein, and that this difference becomes relevant only on difficult targets, such as primary HSC/HPC. In any case, our observation supports the notion that infectious titers (or MOIs) ideally need to be determined on actual target cells because distinct vectors can differ in their transduction capability of defined targets [47,48]. Importantly, independent of the used vector, genemarked cells contributed significantly to the reconstitution of hematopoiesis at all time points, i.e., there was no evidence for clonal outgrowth of transduced cells. Gene marking, as measured based on eGFP expression, was stable over time for both vector types (compare Fig. 2B). This indicates that neither of the vectors underwent silencing and that none of the gene-marked cells were malignantly transformed. However, detection of both undesired effects can require longer observation periods and/or serial transplantation [2,3,12]. Despite the relatively limited number of animals in each experimental group, we made some interesting observations. First, in both vector groups, mean numbers of clones detectable in the peripheral blood remained essentially constant during the whole follow-up time. This observation was in line with FACS data indicating constant marking levels in the peripheral blood. Together these results suggest that long-term reconstituting HSC were

K. Cornils et al./ Experimental Hematology 2013;41:28–38

successfully transduced with both vectors. Second, at all time points, mean numbers of clones contributing to hematopoiesis per 100,000 transplanted gene-marked cells were approximately two times higher in the Lenti-SF group. Based on such normalization, it might be supposed that the lentiviral vectors were relatively more efficient in transducing HSC/HPC. Notably, transduction conditions were identical for both vector types; the observed disparity apparently reflected an intrinsic vector feature of the lentiviral vector. One potential explanation would refer to the postentry restriction noted for the lentiviral vector in this study, which might be relatively less pronounced in longterm reconstituting HSC. Finally, at later time points of our study, clones harboring insertions nearby RTCGD genes were significantly over-represented for the g-retroviral as compared with the lentiviral vector group. This finding is in good agreement with previous observations [3,5,20].

Conclusions Altogether we have shown here that the used transplantation model in conjunction with highly sensitive detection methods, namely LAM-PCR followed by 454 NGS, allows establishment of a representative picture of hematopoietic reconstitution and assessment of the impact of the used marking vector on the clonal repertoire. Obviously, this model system can, in principle, also be used to investigate further parameters of gene transfer, such as target cell population or transduction conditions [18,19]. At the same time, our qPCR data has clearly indicated that the reconstitution picture obtained with this approach is not complete. Despite the fact that we performed two independent LAMPCRs using two different restriction sites for each sample, some clones contributing to hematopoiesis were not adequately quantified based on 454 sequencing. This is in line with previous data [24] and underlines the need to use improved protocols, such as multi-arm LAM-PCR [23]. Alternatively, novel techniques such as nonrestrictive PCR [32] or barcoded vectors for pure clonality analyses without the information of the vector location [34,46,49– 51] can help to overcome these restrictions.

Funding disclosure This work was supported by the Deutsche Forschungsgemeinschaft (DFG; FE568/11-1 to BF, DFG-SPP1230: RO3500/2-1 to IR & SCHM2134/1-2 to MS) and the German Ministry for Research and Education (BMBF; iGene to MS & BF, BMBF-FKZ 0315452 to IR).

Acknowledgments The authors thank G€okhan Arman-Kalcek, Melanie Lachmann, Regine Thiele und Ina Kutschera for excellent technical assistance

37

and Ute Modlich for critical reading of the manuscript. We are grateful to Martijn H. Brugman and Stefan Bartels for kindly providing the ISA tool. Flow cytometry work was performed in the FACS Sorting Core Unit of the UMC Hamburg-Eppendorf.

Conflict of interest disclosure No financial interest/relationships with financial interest relating to the topic of this article have been declared.

References 1. Jordan CT, Lemischka IR. Clonal and systemic analysis of long-term hematopoiesis in the mouse. Genes Dev. 1990;4:220–232. 2. Li Z, D€ullmann J, Schiedlmeier B, et al. Murine leukemia induced by retroviral gene marking. Science. 2002;296:497. 3. Kustikova O, Fehse B, Modlich U, et al. Clonal dominance of hematopoietic stem cells triggered by retroviral gene marking. Science. 2005;308:1171–1174. 4. Enssle J, Trobridge GD, Keyser KA, et al. Stable marking and transgene expression without progression to monoclonality in canine longterm hematopoietic repopulating cells transduced with lentiviral vectors. Hum Gene Ther. 2010;21:397–403. 5. Calmels B, Ferguson C, Laukkanen MO, et al. Recurrent retroviral vector integration at the Mds1/Evi1 locus in nonhuman primate hematopoietic cells. Blood. 2005;106:2530–2533. 6. Tey SK, Brenner MK. The continuing contribution of gene marking to cell and gene therapy. Mol Ther. 2007;15:666–676. 7. Barese CN, Dunbar CE. Contributions of gene marking to cell and gene therapies. Hum Gene Ther. 2011;22:659–668. 8. Hacein-Bey-Abina S, Von Kalle C, Schmidt M, et al. LMO2-associated clonal T cell proliferation in two patients after gene therapy for SCID-X1. Science. 2003;302:415–419. 9. Hacein-Bey-Abina S, von Kalle C, Schmidt M, et al. A serious adverse event after successful gene therapy for X-linked severe combined immunodeficiency. N Engl J Med. 2003;348:255–256. 10. Hacein-Bey-Abina S, Garrigue A, Wang GP, et al. Insertional oncogenesis in 4 patients after retrovirus-mediated gene therapy of SCID-X1. J Clin Invest. 2008;118:3132–3142. 11. Howe SJ, Mansour MR, Schwarzwaelder K, et al. Insertional mutagenesis combined with acquired somatic mutations causes leukemogenesis following gene therapy of SCID-X1 patients. J Clin Invest. 2008;118:3143–3150. 12. Stein S, Ott MG, Schultze-Strasser S, et al. Genomic instability and myelodysplasia with monosomy 7 consequent to EVI1activation after gene therapy for chronic granulomatous disease. Nat Med. 2010;16:198–204. 13. Wang GP, Garrigue A, Ciuffi A, et al. DNA bar coding and pyrosequencing to analyze adverse events in therapeutic gene transfer. Nucleic Acids Res. 2008;36:e49. 14. Cartier N, Hacein-Bey-Abina S, Bartholomae CC, et al. Hematopoietic stem cell gene therapy with a lentiviral vector in X-linked adrenoleukodystrophy. Science. 2009;326:818–823. 15. Cavazzana-Calvo M, Payen E, Negre O, et al. Transfusion independence and HMGA2 activation after gene therapy of human b-thalassaemia. Nature. 2010;467:318–322. 16. Boztug K, Schmidt M, Schwarzer A, et al. Stem-cell gene therapy for the Wiskott-Aldrich syndrome. N Engl J Med. 2010;363:1918–1927. 17. Biffi A, Bartolomae CC, Cesana D, et al. Lentiviral vector common integration sites in preclinical models and a clinical trial reflect a benign integration bias and not oncogenic selection. Blood. 2011; 117:5332–5339. 18. Kustikova OS, Schiedlmeier B, Brugman MH, et al. Cell-intrinsic and vector-related properties cooperate to determine the incidence and consequences of insertional mutagenesis. Mol Ther. 2009;17:1537–1547.

38

K. Cornils et al./ Experimental Hematology 2013;41:28–38

19. Maetzig T, Brugman MH, Bartels S, et al. Polyclonal fluctuation of lentiviral vector-transduced and expanded murine hematopoietic stem cells. Blood. 2011;117:3053–3064. 20. Kustikova OS, Geiger H, Li Z, et al. Retroviral vector insertion sites associated with dominant hematopoietic clones mark ‘‘stemness’’ pathways. Blood. 2007;109:1897–1907. 21. Kustikova OS, Modlich U, Fehse B. Retroviral insertion site analysis in dominant haematopoietic clones. Methods Mol Biol. 2009;506: 373–390. 22. Schmidt M, Schwarzwaelder K, Bartholomae C, et al. High-resolution insertion-site analysis by linear amplification-mediated PCR (LAMPCR). Nat Methods. 2007;4:1051–1057. 23. Harkey MA, Kaul R, Jacobs MA, et al. Multiarm high-throughput integration site detection: limitations of LAM-PCR technology and optimization for clonal analysis. Stem Cells Dev. 2007;16:381–392. 24. Gabriel R, Eckenberg R, Paruzynski A, et al. Comprehensive genomic access to vector integration in clinical gene therapy. Nat Med. 2009; 15:1431–1436. 25. Brady T, Roth SL, Malani N, et al. A method to sequence and quantify DNA integration for monitoring outcome in gene therapy. Nucleic Acids Res. 2011;39:e72. 26. Li Z, Schwieger M, Lange C, et al. Predictable and efficient retroviral gene transfer into murine bone marrow repopulating cells using a defined vector dose. Exp Hematol. 2003;31:1206–1214. 27. Cornils K, Lange C, Schambach A, et al. Stem cell marking with promotor-deprived self-inactivating retroviral vectors does not lead to induced clonal imbalance. Mol Ther. 2009;17:131–143. 28. Weber K, Bartsch U, Stocking C, Fehse B. A multicolor panel of novel lentiviral ‘‘gene ontology’’ (LeGO) vectors for functional gene analysis. Mol Ther. 2008;16:698–706. 29. Zufferey R, Dull T, Mandel RJ, et al. Self-inactivating lentivirus vector for safe and efficient in vivo gene delivery. J Virol. 1998;72:9873–9880. 30. Schambach A, Bohne J, Baum C, et al. Woodchuck hepatitis virus post-transcriptional regulatory element deleted from X protein and promoter sequences enhances retroviral vector titer and expression. Gene Ther. 2006;13:641–645. 31. Schambach A, Galla M, Modlich U, et al. Lentiviral vectors pseudotyped with murine ecotropic envelope: increased biosafety and convenience in preclinical research. Exp Hematol. 2006;34:588–592. 32. Paruzynski A, Arens A, Gabriel R, et al. Genome-wide highthroughput integrome analyses by nrLAM-PCR and next-generation sequencing. Nat Protoc. 2010;5:1379–1395. 33. Hoffmann S, Otto C, Kurtz S, et al. Fast mapping of short sequences with mismatches, insertions and deletions using index structures. PLoS Comput Biol. 2009;5:e1000502. 34. Bystrykh LV, Verovskaya E, Zwart E, Broekhuis M, de Haan G. Counting stem cells: methodological constraints. Nat Methods. 2012;96:567–574.

35. Edgar RC. Search and clustering orders of magnitude faster than BLAST. Bioinformatics. 2010;26:2460–2461. 36. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 1994;22:4673–4680. 37. Akagi K, Suzuki T, Stephens RM, Jenkins NA, Copeland NG. RTCGD: retroviral tagged cancer gene database. Nucleic Acids Res. 2004;32(Database issue):D523–527. 38. Kustikova OS, Baum C, Fehse B. Retroviral integration site analysis in hematopoietic stem cells. Methods Mol Biol. 2008;430:255–267. 39. Bozorgmehr F, Laufs S, Sellers SE, et al. No evidence of clonal dominance in primates up to 4 years following transplantation of multidrug resistance 1 retrovirally transduced long-term repopulating cells. Stem Cells. 2007;25:2610–2618. 40. Kustikova OS, Wahlers A, Kuhlcke K, et al. Dose finding with retroviral vectors: correlation of retroviral vector copy numbers in single cells with gene transfer efficiency in a cell population. Blood. 2003; 102:3934–3937. 41. Cattoglio C, Maruggi G, Bartholomae C, et al. High-definition mapping of retroviral integration sites defines the fate of allogeneic T cells after donor lymphocyte infusion. PLoS One. 2010;5:e15688. 42. Roeder I, Horn K, Sieburg HB, Cho R, Muller-Sieburg C, Loeffler M. Characterization and quantification of clonal heterogeneity among hematopoietic stem cells: a model-based approach. Blood. 2008;112: 4874–4883. 43. Drize N, Chertkov J, Sadovnikova E, Tiessen S, Zander A. Long-term maintenance of hematopoiesis in irradiated mice by retrovirally transduced peripheral blood stem cells. Blood. 1997;89:1811–1817. 44. Gerrits A, Dykstra B, Kalmykowa OJ, et al. Cellular barcoding tool for clonal analysis in the hematopoietic system. Blood. 2010;115:2610– 2618. 45. Bieniasz PD, Cullen BR. Multiple blocks to human immunodeficiency virus type 1 replication in rodent cells. J Virol. 2000;74:9868–9877. 46. Tervo HM, Goffinet C, Keppler OT. Mouse T-cells restrict replication of human immunodeficiency virus at the level of integration. Retrovirology. 2008;5:58. 47. Shabram P, Aguilar-Cordova E. Multiplicity of infection/multiplicity of confusion. Mol Ther. 2000;2:420–421. 48. Fehse B, Kustikova OS, Bubenheim M, Baum C. Pois(s)ondit’s a question of dose. Gene Ther. 2004;11:879–881. 49. Schepers K, Swart E, van Heijst JW, et al. Dissecting T cell lineage relationships by cellular barcoding. J Exp Med. 2008;205:2309–2318. 50. Lu R, Neff NF, Quake SR, Weissman IL. Tracking single hematopoietic stem cells in vivo using high-throughput sequencing in conjunction with viral genetic barcoding. Nat Biotechnol. 2011;29:928–933. 51. Glimm H, Ball CR, von Kalle C. You can count on this: barcoded hematopoietic stem cells. Cell Stem Cell. 2011;9:390–392.

K. Cornils et al./ Experimental Hematology 2013;41:28–38 Supplementary Table E1. GFP-positive cells of different blood lineages at final analysis Lenti-SF

Gr-1b CD11bc B220d CD3e

Meana

SD

Mean

SD

28.0 33.5 11.6 13.5

31.62 41.20 10.50 15.90

61.4 60.0 60.4 63.9

31.53 43.82 42.91 27.71

SD 5 standard deviation. Mean of all 4 animals. b Marker for granulocytes. c marker for monocytes. d B-cell marker. e T-cell marker. a

SF91

38.e1

Recipienta

Cluster IDb

Gene symbol

SF91

I.1.2

1.2combi4 1.2combi43 1.2combi52 1.2combi59 1.2combi79 1.2combi9 1.4combi15 1.4combi21 1.4combi34 1.4combi37 1.4combi42 1.4combi51 1.4combi73 1.4combi8 1.4combi93 1.5combi11 1.5combi16 1.5combi18 1.5combi19 1.5combi3 1.5combi32 1.5combi33 1.5combi39 1.5combi40 1.5combi46 1.5combi48 1.5combi49 1.5combi55 1.5combi75 1.5combi9 1.6combi1 1.6combi14 1.6combi19 1.6combi24 1.6combi29 1.6combi30 1.6combi31 1.6combi34 1.6combi36 1.6combi44 1.6combi48 1.6combi58 1.6combi6 1.6combi7 1.6combi73

A430084P05Rik Il12a Mllt3 Lnpep Auts2 Prkce Ddit4 Mllt6 Ppcdc Ifngr1 Hdac9 Lasp1 Ninj2 Rcor1 Ahdc1 Ccr7 Rftn1 Cebpb Sorl1 Edg1 Pcbp1 Bach2 Ahi1 Prcp Evi1 4930511I11Rik Hoxa9 Skap2 Wasf2 Arhgef2 Evi1 Lyz Edg1 Ifi47 Tpd52 Col15a1 Smap1l Ccr7 Phtf2 Lyst Smarcal1 Smyd3 Slc12a3 Dok3 Cspg4

I.1.4

I.1.5

I.1.6

Chromc 11 B3 3 E1 4 C4 17 A2-3 5 G2 17 E4 10 B3 11 D 9B 10 A3 12 A3 11 C-D 6 F1 12 F1 4 D2-3 11 D 17 C 2 H3 9 A5 3 G1 6 D1 4 A5 10 A3 7 E2 3 A3 17 A3.3 6 B3 6 B3 4 D2.3 3 F1 3 A3 10 D2 3 G1 11 B1.2 3 A1-A2 4 B1-B3 4 D2.2 11 D 5 A3 13 A1 1 C3 1 H3 8 C5 13 B1 9B

Definition or (proposed) function

Hitsd

RIKEN cDNAA430084P05 gene Interleukin 12a Myeloid/Lymphoid or mixed lineage-leukemia translocation to 3 homolog (Drosophila) Leucyl/Cystinyl aminopeptidase Autism susceptibility candidate 2 Protein kinase C, epsilon DNA-damage-inducible transcript 4 Myeloid/lymphoid or mixed lineage-leukemia translocation to 6 homolog (Drosophila) Phosphopantothenoyl cysteine decarboxylase Interferon gamma receptor 1 Histone deacetylase9 LIM and SH3 protein1 Ninjurin 2 REST corepressor 1 AThook, DNA binding motif, containing 1 Chemokine (C-Cmotif) receptor 7 Raftlin lipid raft linker 1 CCAAT/enhancer binding protein (C/EBP), beta Sortilin-related receptor, LDLR class A repeats-containing Endothelial differentiation sphingolipid G-protein-coupled receptor 1 Poly(rC)binding protein 1 BTB and CNC homology 2 Abelson helper integration site j Myeloblastosisoncogene Prolylcarboxypeptidase (angiotensinaseC) Ecotropic viral integration site 1 RIKEN cDNA4930511I11 gene HomeoboxA7 j HomeoboxA9 Src family associated phosphoprotein 2 WAS protein family, member 2 Rho/Rac guanine nucleotide exchangefactor (GEF) 2 Ecotropic viral integration site 1 Lysozyme Endothelial differentiation sphingolipid G-protein–coupled receptor 1 Interferon gamma inducible protein 47 Tumor protein D52 Procollagen, typeXV Stromal membrane-associated protein 1–like Chemokine (C-Cmotif) receptor 7 Putative homeodomain transcription factor 2 Lysosomal trafficking regulator Swi/SNF-related matrix-associated, actin-dependent regulator of chromatin, subfamily a-like 1 SET and MYND domain containing 3 Solute carrier family 12, member 3 Docking protein 3 Chondroitin sulfate proteoglycan 4

1 3 2 5 1 2 2 2 2 2 3 2 4 1 1 8 1 25 2 4 1 14 61 1 25 3 25 2 4 2 25 1 4 3 1 1 1 8 1 2 1 1 1 1 3 (continued)

K. Cornils et al./ Experimental Hematology 2013;41:28–38

Vector

38.e2

Supplementary Table E2. List of closest gene to integration site registered in the RTCGD

Supplementary Table E2. (continued ) Vector

Lenti-SF

Recipienta

I.4.2

I.4.4 I.4.5

Gene symbol

Chromc

1.6combi8 1.6combi80 1.6combi81 4.2combi_10 4.2combi_20 4.2combi_33 4.4combi_41 4.4combi_7 4.5combi_20 4.5combi_22 4.5combi_33 4.5combi_34 4.5combi_9 4.7combi_25

Hectd2 Rgag4 Sp2 Zcchc6 Elovl5 Pdcd6ip Ctsc Mlstd2 Zscan22 Parp11 2310022M17Rik Pscd1 Mlstd2 Atf3

19 C2 XD 11 D 13 B2 9 E1 9 F2 7 D3-E1.1 7 F2 7 A1 6 F3 11 A4 11 E2 7 F2 1 H6

Definition or (proposed) function HECT domain containing 2 Retrotransposon gag domain containing 4 Sp2 transcription factor zincfinger, CCHC domain containing 6 ELOVL family member 5, elongation of long chain fatty acids (yeast) Programmed cell death 6 interacting protein Cathepsin C Male sterility domain containing 2 Zincfinger and SCAN domain containing 22 Poly(ADP-ribose)polymerase family, member 11 RIKEN cDNA2310022M17 gene Pleckstrin homology, Sec7 and coiled-coil domains 1 Male sterility domain containing 2 Activating transcription factor 3

Hitsd 1 2 1 1 3 1 2 1 1 2 1 1 1 1

Recipient: Transplanted mice are numbered: I represents cohort, 1 represents the group (1 5 SF91, 4 5 Lenti-SF) and the last number resembles the individual mouse. The consensus sequence after processing the 454-sequencing data was blasted against the mouse genome via the ISA tool. Genes in a window of 6250 kb around the insertion site were analyzed. The closest gene to the insertion site was checked against the RTCGD database [37]. c Integration site of vector on chromosome. d Number of hits in the RTCGD. a

b

K. Cornils et al./ Experimental Hematology 2013;41:28–38

I.4.7

Cluster IDb

38.e3