Proc. Natl. Acad. Sci. USA
`Vol. 92, pp. 6728-6732, July 1995
`Genetics
`
`Generation of a high-titer retroviral vector capable of expressing
`high levels of the human (8-globin gene
`(hemoglobinopathies/gene therapy/locus control region)
`
`MICHEL SADELAIN*t, C. H. JASON WANG*, MICHAEL ANTONIOUt, FRANK GROSVELDt§,
`AND RICHARD C. MULLIGAN*¶
`*Whitehead Institute for Biomedical Research and Department of Biology, Massachusetts Institute of Technology, Nine Cambridge Center, Cambridge, MA
`02142-1479; and $Laboratory for Gene Structure and Expression, Medical Research Council National Institute for Medical Research, The Ridgeway,
`Mill Hill, London, NW7, United Kingdom
`Communicated by David E. Housman, Massachusetts Institute of Technology, Cambridge, MA, December 12, 1994
`
`Retrovirus-mediated gene transfer into he-
`ABSTRACT
`matopoietic cells may provide a means of treating both
`inherited and acquired diseases involving hematopoietic cells.
`Implementation of this approach for disorders resulting from
`mutations affecting the 8-globin gene (e.g., P-thalassemia and
`sickle cell anemia), however, has been hampered by the
`inability to generate recombinant viruses able to efficiently
`and faithfully transmit the necessary sequences for appropri-
`ate gene expression. We have addressed this problem by
`carefully examining the interactions between retroviral and
`j3-globin gene sequences which affect vector transmission,
`stability, and expression. First, we examined the transmission
`properties ofa large number ofdifferent recombinant proviral
`genomes which vary both in the precise nature of vector,
`18-globin structural gene, and locus control region (LCR) core
`sequences incorporated and in the placement and orientation
`of those sequences. Through this analysis, we identified one
`specific vector, termed Mj36L, which carries both the human
`,8-globin gene and core elements HS2, HS3, and HS4 from the
`LCR and faithfully transmits recombinant proviral sequences
`to cells with titers greater than 106 per ml. Populations of
`murine erythroleukemia (MEL) cells transduced by this virus
`expressed levels of human j8-globin transcript which, on a per
`gene copy basis, were 78% of the levels detected in an
`MEL-derived cell line, Hull, which carries human chromo-
`some 11, the site of the 13-globin locus. Analysis of individual
`transduced MEL cell clones, however, indicated that, while
`expression was detected in every clone tested (n = 17), the
`levels of human j8-globin treatment varied between 4% and
`146% of the levels in Hull. This clonal variation in expression
`levels suggests that small ,-globin LCR sequences may not
`provide for as strict chromosomal position-independent ex-
`pression of 18-globin as previously suspected, at least in the
`context of retrovirus-mediated gene transfer.
`
`The successful treatment of f3-globin disorders by gene therapy
`will likely require the efficient and stable introduction of a
`functional ,3-globin gene into self-renewing hematopoietic
`stem cells. While we and others had established that retroviral
`vectors encoding the human f-globin structural gene could
`3-globin in mice
`direct the erythroid-specific synthesis of
`engrafted with genetically modified bone marrow cells (1-4),
`these studies suffered from two important limitations. First,
`the retroviral titers generally obtained with those vectors made
`it difficult to generate long-term bone marrow chimeras
`engrafted with efficiently transduced cells. A second problem
`was the variable and generally low level of human 13-globin
`expression in vivo observed in the reconstituted animals.
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked "advertisement" in
`accordance with 18 U.S.C. §1734 solely to indicate this fact.
`
`The prospects for increasing the level of f-globin expression
`obtainable in retroviral vectors were dramatically improved by
`the recent finding that DNA sequences located considerable
`distances from the human l3-globin structural gene on chro-
`mosome 11, termed locus control region (LCR) sequences (5,
`6), play a critical role in the transcriptional control of genes
`within the P-globin-like gene cluster (reviewed in refs. 7-9).
`Specifically, transgenic experiments from a number of labo-
`ratories have indicated that the linkage of LCR sequences to
`the human 13-globin structural gene dramatically increases the
`level of expression of 3-globin observed in erythroid cells
`(10-13). Some studies further suggested that LCR sequences
`may differ from classical transcriptional enhancer sequences in
`that they provide for the chromosomal position-independent
`expression of linked genes (10-16). Smaller fragments, more
`suitable for insertion into vectors, have recently been defined,
`which have been reported to retain partial LCR activity
`(17-29). Recent efforts by several groups to incorporate LCR
`subfragments into ,3-globin retroviral vectors, however, have
`resulted in vector rearrangements (ref. 30 and our observa-
`tions), poor titers (31), or very low expression of the ,B-globin
`gene (32).
`We report here a systematic study of the ,B-globin gene, LCR
`core sites, and retroviral sequences that control vector trans-
`mission. This analysis has led to the generation of a high-titer,
`genomically stable retroviral vector bearing the human ,3-glo-
`bin gene and the LCR core sites HS2, HS3, and HS4. This
`vector confers elevated and erythroid-specific expression of
`the ,B-globin gene. Expression in different clones of transduced
`murine erythroleukemia (MEL) cells is variable, however,
`raising questions about the ability of small LCR sequences to
`confer position-independent gene expression, at least in the
`context of a retroviral vector.
`
`MATERIALS AND METHODS
`Generation of Vectors and Packaging Cell Lines. The
`retroviral vectors pSG and pMFG are described elsewhere (33,
`34). The 13-globin gene was subcloned from pSVX-Neo (1).
`The f3-globin coding sequence (13CS) consists of a 444-bp Nco
`I-linker-containing ,-globin sequence from the translational
`start to the stop codon derived from the cDNA. The HS2, HS3,
`and HS4 fragments (see text and refs. 19-22 and 27-29) were,
`respectively, given HindIII, EcoRI, and Mlu I linkers and
`subcloned in all possible orientations in a Sac II-HindIlI-
`
`Abbreviations: LCR, locus control region; LTR, long terminal repeat;
`MEL, murine erythroleukemia; MLV, murine leukemia virus.
`tPresent address: Department of Human Genetics, Memorial Sloan-
`Kettering Cancer Center, Box 182,1275 York Avenue, New York, NY
`10021.
`§Present address: Department of Cell Biology, Erasmus University,
`Postbus 1738, 3000 Dr Rotterdam, The Netherlands.
`1To whom reprint requests should be addressed.
`
`6728
`
`Downloaded from https://www.pnas.org by 163.116.135.92 on January 21, 2023 from IP address 163.116.135.92.
`
`SKI Exhibit 2024
`Page 1 of 5
`
`

`

`Genetics: Sadelain et al.
`
`Proc. Natl. Acad. Sci. USA 92 (1995)
`
`6729
`
`EcoRI-Mlu I-Sac II polylinker. The vectors described herein,
`lacking any selectable marker, were introduced by calcium
`phosphate cotransfection with pSV2Neo (35) into the qi-CRE
`packaging cell line (36). After 10 days of selection in G418
`(Geneticin, Sigma) at 1 mg/ml, the G418-resistant colonies
`(>50 per transfection) were pooled by trypsinization and
`grown in Dulbecco's modified Eagle's medium supplemented
`with 10% calf serum, penicillin, and streptomycin without
`G418.
`Fibroblast Infection and Viral Transmission Assay. Five
`milliliters of supernatant harvested from pooled producer cells
`was applied onto 106 NIH 3T3 fibroblasts. Polybrene (Sigma)
`was added at 8 ,ug/ml. After 4 hr, the conditioned medium was
`removed. Three days later, genomic DNA was extracted from
`the confluent NIH 3T3 fibroblasts. For Southern analysis,
`10-,ug samples of genomic DNA from the producer and the
`corresponding 3T3 target cells were digested overnight with
`the restriction enzyme Nhe I [which cuts in both long terminal
`repeats (LTRs)] and run side by side on a 1% agarose gel.
`[32P]dCTP-labeled 13-globin probes were generated by using
`the human Nco I-EcoRI genomic template. Band intensity was
`quantified by using a phosphorimager (Fuji Bio-Imaging).
`Vector signal intensity was normalized to that of the endog-
`enous band. Very even loading in Figs. 1-3 allows for direct
`reading of the gels as shown.
`MEL Cell Infection and Screening. C88 MEL cells were
`grown in RPMI medium/10% fetal bovine serum/penicillin/
`streptomycin. Infection was performed by a 48-hr cocultivation
`of 105 MEL cells, treated for the previous 18 hr with tunica-
`mycin (Sigma) at 0.2 ,ug/ml, on the ecotropic producer in the
`presence of Polybrene at 4 ,ug/ml. After cocultivation, MEL
`cells were subcloned onto 96-well plates at 0.2 cell per well.
`Single clones were then expanded and genomic DNA was
`prepared. Infected clones were screened by PCR using the
`primers GCAAGAAAGTGCTCGGTG in exon 2 and TCT-
`GATAGGCAGCCTGCA in exon 3. The DNA of positive
`clones was further analyzed by Southern blotting for copy
`number and integration site.
`MEL Cell Induction and Globin mRNA Quantification.
`After 4 days in 2% (vol/vol) dimethyl sulfoxide, RNA was
`extracted from each MEL cell clone by using the lithium
`chloride/urea procedure (37). This induction-extraction pro-
`cedure was repeated at least once. RNA was quantified by
`RNase protection (38) using probes described in ref. 39 for
`human 13-globin and a 180-nucleotide Pst I-BamHI genomic
`fragment (40) for mouse a-globin. One to 2 ,g of RNA per
`extract was incubated with both admixed probes, which, after
`RNase digestion, were electrophoresed on a 6% polyacryl-
`amide/urea gel. Signal intensity for each band (the a-globin
`serving as internal control for both induction and loading onto
`the gel) was determined by using the phosphorimager.
`
`RESULTS
`Features of Retroviral Vector Design Which Affect Trans-
`mission of the Human 18-Globin Structural Gene. To assess the
`titer and genomic stability of different retroviral vectors
`bearing the human ,3-globin gene, we first developed a direct
`assay to characterize and quantitate vector transmission. This
`assay is based on a quantitative Southern blot analysis of vector
`copy number in both transfected packaging cells and infected
`3T3 fibroblasts. Supernatant is harvested from a pool of stable
`virus-producing cell clones and used to infect the target
`fibroblasts under defined conditions (see Materials and Meth-
`ods). Southern blot analysis of vector DNA copy number in the
`polyclonal producer and target cells allows for a direct eval-
`uation of average transmission for any retroviral construct.
`This is achieved by direct comparison of signal intensity in both
`genomic DNA extracts run side by side and therefore does not
`rely on indirect readouts dependent on gene expression (such
`
`as G418 resistance). This method provides not only for a direct
`measurement of gene transfer but also for the detection of
`rearrangements of any given construction. As shown in Fig. 1,
`transmission of a minimal 1.8-kb 13-globin gene and promoter
`(-129 to + 1650) cloned in reverse orientation in the pSG
`vector (33) was not detectable in this assay (lanes A). Trans-
`mission of the larger 2.8-kb Sph I-Pst I (-615 to +2163) globin
`fragment was also not detectable (data not shown). In contrast,
`the f3-globin coding sequence alone transmitted very well,
`showing comparable signals in the packaging and target cell
`DNAs (lanes B). The high titer of the latter vector, however,
`was greatly reduced by inclusion of the core sites HS2, HS3,
`and HS4 of the LCR (see below), which were introduced in the
`vector upstream of the globin sequence in either orientation
`(lanes C and D).
`To determine which P-globin genomic sequences were re-
`sponsible for the dramatic reduction in titer, we constructed a
`set of recombinant vectors bearing systematic deletions of
`untranslated regions of the gene. As shown in Fig. 2, deletion
`of the 5' untranslated region and intron 1, either alone or in
`combination, failed to increase viral titers (lanes A-D). De-
`letion of intron 2 (lanes E) increased transmission to levels
`similar to the coding sequence vector (lanes I), yet further
`combined deletions did not increase titer any more (lanes
`F-H). Because ,B-globin DNA templates lacking intron 2 direct
`poor ,B-globin expression (refs. 24 and 41-44 and data not
`shown), we examined transmission of two subintronic dele-
`tions (from +518 to + 1288 and +679 to + 953). Both yielded
`similar titers, which were still 10-fold down from the intronless
`gene (shown for the first deletion, lanes J, Fig. 2). Therefore,
`intron 2 alone accounts for reduced titers of 13-globin gene, and
`deletions within intron 2, while increasing titer, still transmit
`less well than intronless constructions.
`Recombinant Human 3-Globin Genes Yield Higher Titers
`in the pMFG Vector. Since decreased viral transmission may
`result from a negative interaction between 13-globin and ret-
`roviral sequences, which is only partially alleviated by intron 2
`subdeletions, we compared transmission of /3-globin sequences
`in different vectors. Modifications of the length of gag se-
`quence, replacement of the Moloney murine leukemia virus
`(MLV) packaging signal by that of Ha-Ras MLV, and muta-
`tion of the splice donor site present in pSG failed to increase
`vector transmission (data not shown). Subcloning recombinant
`,3-globin fragments into the pMFG vector (34), however,
`significantly increased transmission (Fig. 3). Lanes A and D
`show transmission of the full P-globin fragment; lanes B and
`E, the genomic sequence bearing the intron 2 deletion shown
`
`0.51.0
`
`A
`
`B
`
`C
`
`D
`
`Comparison of vector transmission by Southern blot anal-
`FIG. 1.
`ysis of vector copy number in the packaging cell (left lane) and target
`cell (right lane) genomic DNA. In lanes A, 1.8-kb 13-globin gene and
`promoter; in lanes B, f3CS; in lanes C and D, f3CS and the three core
`elements HS2, HS3, and HS4 in sense and antisense orientation,
`respectively. See text for sequence description. Genomic DNA was
`3-globin probe.
`digested with Nhe I and blots were probed with a
`Control lanes show signal intensity for one proviral copy per cell (1.0)
`and one copy per two cells (0.5). Even loading (data not shown) allows
`for directly comparing signal intensity between lanes.
`
`Downloaded from https://www.pnas.org by 163.116.135.92 on January 21, 2023 from IP address 163.116.135.92.
`
`SKI Exhibit 2024
`Page 2 of 5
`
`

`

`6730
`
`Genetics: Sadelain et al.
`
`Proc. Natl. Acad. Sci. USA 92 (1995)
`
`0510 A
`
`ri
`
`B
`
`E
`D
`C
`ii I i r
`.
`.S5m
`
`F
`
`r-
`
`H
`G
`1!
`. ..5S-inait_ v'/2'
`
`A
`m
`
`B
`mI
`
`C
`F
`E
`D
`H
`G
`o O
`Im---- mm mm ol-- r(e
`.....,.. .,'.IW
`
`ad
`
`Effect of ,B-globin gene untranslated sequences on vector
`FIG. 2.
`transmission. Southern blot analysis of packaging cell (iI-CRE, left
`lane) and target cell (NIH 3T3, right lane) genomic DNA. Lanes: A,
`1.8-kb 13-globin gene and promoter; B, 5' untranslated region deletion;
`C, intron 1 deletion; D, B and C combined deletions; E, intron 2
`deletion; F, B and E combined deletions; G, C and E combined
`3-globin coding
`deletions; H, B, C, and E combined deletions; I,
`sequence only; J, 770-bp intron 2 deletion. See Fig. 1 legend and text
`for sequences and methods.
`
`in lanes J of Fig. 2; and lanes C and F, the globin coding
`sequence, in the vectors pSG and pMFG, respectively. Com-
`parison between lanes B and E shows a gain in titer of about
`5-fold when the pMFG vector is used, showing that altering
`vector sequences can affect transmission.
`The Core Sites of HS2, HS3, and HS4 Can Be Incorporated
`into a High-Titer Genomically Stable Retroviral Vector. Pre-
`vious studies of LCR sequences had indicated that a "minilo-
`cus" approximately 20 kb in size, comprising sites HS1, HS2,
`HS3, and HS4, possessed close to complete LCR activity (10,
`11, 14). However, because of the inability to transmit se-
`quences of such size in retrovirus vectors, we chose to attempt
`to incorporate "core elements" of HS2, HS3, and HS4, which
`are markedly reduced in size and have been shown to retain at
`least partial LCR activity, into vectors carrying the (3-globin
`structural gene. The following sequences were employed for
`our studies, based on studies in transgenic animals (17-22,
`27-29): (i) the 283-bp Sac I-Ava I core HS4 fragment (29); (ii)
`the 260-bp core HS3, including the Hph I-Fnu4HI segment
`(27, 28) and the upstream NF-E2 site, which encompasses all
`sites footprinted in vivo (45, 46); and (iii) a 478-bp HindIII-
`SnaBI HS2 fragment encompassing all sites footprinted in vivo
`(45, 47) in the core HindIII-Xba I fragment (19-22) and most
`of the adjacent alternating purine-pyrimidine stretch (21, 48).
`As shown previously in Fig. 1, one specific orientation of the
`sites greatly reduced proviral transmission. Accordingly, the
`three sites were subcloned in such a way that all eight permu-
`tations of relative arrangement were generated. The resulting
`cassettes were introduced, in sense and antisense orientations
`upstream of the ,B-globin sequence in the high-titer f3-globin
`vector shown in lanes B of Fig. 1. As shown in Fig. 4,
`transmission of these vectors shows significant variations with
`regard to copy number and genomic stability (e.g., the two
`species transmitted in lanes H). Among these 16 constructs, we
`identified four combinations which were transmitted at high
`titer without rearrangements (lanes A, E, L, and N). All bear
`HS2 in the antisense orientation relative to retroviral tran-
`1o0 A
`C
`B
`D
`E
`F
`
`Comparison of transmission of 3-globin sequences in two
`FIG. 3.
`different vectors. Lanes A-C, pSG vector; lanes D-F, pMFG vector.
`Lanes A and D, 3-globin gene (as in Fig. 1, lanes A); lanes B and E,
`intron 2 deleted minigene (as in Fig. 2, lanes J); lanes C and F, ,BCS
`(as in Fig. 1, lanes B). See Figs. 1 and 2 for nomenclature and methods.
`Signal intensity between lanes can be directly compared owing to very
`even loading (data not shown) except for lanes F, which were loaded
`with 5 ,ug.
`
`I
`
`r
`
`J
`---II
`
`K
`
`L
`~ l
`
`M
`
`N o u
`-- 0
`
`1-
`
`Effect of HS orientation on vector transmission. Southern
`FIG. 4.
`blot analysis of packaging cell (4i-CRE, left lane) and target cell (NIH
`3T3, right lane) genomic DNA (see previous figure legends for
`methods). The top band corresponds to the endogenous globin genes.
`HS orientation (5, sense; s, antisense) in the vector is (in the order
`HS4-HS3-HS2): A, Sss; B, reversed A; C, sss; D, reversed C; E, SSs;
`F, reversed E; G, sSs; H, reversed G; I, SsS; J, reversed I; K, ssS; L,
`reversed K; M, sSS; and N, reversed M. Lane 3T3, NIH 3T3.
`
`scription. The reason for the lack of the endogenous band in
`lanes E is unclear.
`For further studies, we selected the combination with the
`highest titer and in which HS2 and HS3 were oriented in the
`same direction as the ,B-globin gene and adjacent to its
`promoter. The resulting vector, M,B6L, is shown in Fig. 5. In
`summary, this construct contains a 2150-bp SnaBI-PstI ,B-glo-
`bin fragment including a full intron 1 and a 476-bp intron 2 and
`lacking the 3' enhancer [on the basis of other studies that
`suggest it is redundant in the presence of the LCR (24)]. The
`1-kb LCR fragment comprises the HS2, HS3, and HS4 core
`sequences described above (Fig. 5A). Southern blot analysis of
`NIH 3T3 cells transduced by M/36L indicates that approxi-
`mately one proviral copy is transferred per cell (Fig. 5B). On
`the basis of previous comparisons of proviral copy number and
`titers measured by selectable marker expression (unpublished
`results), the titer of M136L is approximately 1 X 106 per ml.
`Mj86L Directs Elevated j8-Globin Expression, Detectable at
`All Integration Sites in MEL Cells. To test the biological
`activity of M136L, we chose MEL cells as target cells for gene
`transfer, as those cells have been extensively employed in
`studies of /3-globin transcription (10, 14-16, 23-26). A panel of
`MEL cell clones infected with M,B6L was generated by cocul-
`tivation of C88 MEL cells and virus-producing cells and
`subsequent subcloning by limiting dilution (see Materials and
`Methods). Southern blot analyses confirmed that intact vector
`copies were integrated at different chromosomal positions for
`each clone (data not shown). Expression of human ,B-globin
`was quantitated by RNase protection using total RNA ex-
`tracted from dimethyl sulfoxide-induced clones. Levels of
`human /3-globin mRNA transcripts were measured by using the
`phosphorimager and normalized to endogenous a- or (3-major
`globin transcripts. This analysis is summarized in Table 1,
`where values are normalized to human B3-globin expression in
`Hull, a MEL cell containing human chromosome 11 under
`selective pressure (49), based on the comparison of their
`respective human , to murine a mRNA ratios (calculations
`based on human f3/murine {3-major yielded very similar re-
`sults). In a pool of 105 unselected MEL cells infected with the
`parent vector lacking the LCR, M(86, the mean level of
`expression was 2.2% of that measured in Hull. Mean expres-
`sion was 78% in MEL cell pools infected with M,B6L. Human
`(-globin transcripts were detected in NIH 3T3 fibroblasts but
`at levels about 1/100 of the level measured from the equal
`amount of RNA from induced MEL cells (Table 1). This
`
`Downloaded from https://www.pnas.org by 163.116.135.92 on January 21, 2023 from IP address 163.116.135.92.
`
`SKI Exhibit 2024
`Page 3 of 5
`
`

`

`Genetics: Sadelain et al.
`
`Proc. Natl. Acad. Sci. USA 92 (1995)
`
`6731
`
`A
`
`Pst
`+2792
`
`intron 2
`deletion
`
`SnaB I
`- 265
`
`enhancer
`deletion
`
`I+
`
`SD
`
`SA
`
`i HS2
`
`HS3
`
`HS4
`
`I
`
`B
`
`0.5
`
`1
`
`M16
`
`I
`
`M,B6L
`IF---l
`
`690
`
`476
`
`423
`
`280
`
`283
`
`5230
`
`I .*14 IF_jj&..
`
`:.
`
`- - --
`
`9-
`
`IB 40
`
`0 ----|
`
`MFG
`
`HUMAN B-GLOBIN GENE
`
`LCR
`
`SG
`
`(A) Schematic representation of the M136L vector. Black boxes represent the ,3-globin exons. Arrows above the HS boxes indicate their
`FIG. 5.
`orientations. Numbers below delineated segments represent fragment length in bp. 4u+, packaging signal. White boxes at ends represent LTRs. See
`text for exact sequences. SD, splice donor; SA, splice acceptor; MFG, 5' sequences derived from the MFG vector; SG, 3' sequences derived from
`the SG vector. (B) Transmission characteristics of vectors M036L and M136, the parent ,B-globin gene vector lacking the LCR fragment. Southern
`blot analysis is carried out as in the previous figures. Length on left is in bp.
`
`finding is consistent with other studies of ,B-globin expression
`in transfected fibroblasts (14, 15, 23, 24). In subcloned MEL
`cells, the expression values ranged from 4% to 146% (ML1-
`ML7, Table 1). The level measured for each clone was very
`stable upon repeated measurement as well as upon reinduc-
`tion. Importantly, in this set of clones and others with slightly
`different ,B-globin gene boundaries (MS and RCM, unpub-
`lished observations), which represent a total of 17 clones,
`human ,B-globin expression, albeit variable in amount, was
`readily detected in all cases. This is in contrast to clones
`infected with the vector lacking LCR sequences (M36), in
`which human f3-globin expression was undetectable in 3 of 4
`clones (M1-M4; Table 1).
`
`Table 1. Human ,B-globin expression in dimethyl sulfoxide-induced
`MEL cells and in NIH 3T3 fibroblasts
`
`Hug3 RNA/Ma
`Hu,B gene
`RNA per vector
`Cell
`copy number
`LCR
`copy, %
`-
`0
`MEL
`0
`Hull
`100
`1
`+
`MEL/M,B6
`-
`0.5
`2.2
`MEL/Mg6L
`78
`1.6
`+
`3T3/M,86
`-
`1.9
`0.2
`3T3/M36L
`1.8
`0.6
`+
`Ml
`1
`2
`-
`M2
`1
`<0.1
`-
`M3
`1
`<0.1
`-
`M4
`1
`<0.1
`-
`MLl
`1
`61
`+
`ML2
`1
`101
`+
`ML3
`1
`146
`+
`ML4
`1
`4
`+
`ML5
`3
`28
`+
`ML6
`4
`+
`35
`ML7
`5
`+
`74
`Protected RNA transcripts were fractionated on a 6% polyacryl-
`amide/urea gel and their radioactivities were measured by using the
`phosphorimager. Results are expressed as human 13-globin mRNA/
`endogenous murine a-globin mRNA divided by vector copy number,
`normalized to the values measured in the Hull clone (see text). Values
`in NIH 3T3 cells are expressed as human ,B-globin RNA divided by
`vector copy number in 1 ,ug of total RNA normalized to the human
`,3-globin signal in 1 ,ug of total RNA from dimethyl sulfoxide-induced
`Hull cells. In the first six lines, data from a large polyclonal cell
`population; in the rest of the table, data from individual MEL cell
`clones.
`
`DISCUSSION
`In this study, we report the resolution of two major obstacles
`to the implementation of retrovirus-based genetic treatment of
`,B-globin disorders: first, the generation of a high-titer retro-
`viral vector suitable for obtaining the efficient transduction of
`hematopoietic stem cells and, second, the generation of vectors
`capable of expressing high levels of the f-globin gene in an
`erythroid-specific fashion. Our systematic study of vector
`transmission indicated problems associated with incorporation
`of both the ,B-globin structural gene itself and the LCR core
`sequences. In the case of the f3-globin structural gene, our
`results parallel those of Miller et al. (43), who had observed
`that the poor transmission of retroviral vectors carrying the
`P-globin structural gene was caused by both untranslated
`sequences in the 5' and 3' regions of the Hpa I-Xba I P3-globin
`fragment and sequences within intron 2. Our results differ
`slightly from those of that study in that we found that se-
`quences within intron 2 alone primarily account for the poor
`transmission of f3-globin vectors (Fig. 2). We found that
`features of retroviral vector design also affected proviral
`transmission of ,3-globin sequences. For example, modifica-
`tions of the vector sequences juxtaposed to the 3' end of the
`13-globin gene indicated that including the Moloney MLV
`splice acceptor region from pMFG (34) resulted in a 5-fold
`increase in titer. Deletion of the U3 region of the 3' LTR (33)
`did not decrease titer and, in fact, led to a moderate (2-fold)
`increase (data not shown).
`We also confirmed the studies of Novak et al. (30) that
`suggested that LCR sequences can lead to significant vector
`instability (data not shown). In our studies, even the core
`fragments incorporated together led to decreased transmission
`and vector instability in particular sequence arrangements
`(Figs. 1 and 4). To overcome this problem, we examined the
`transmission of all 16 permutations of the core sites, and we
`identified combinations that result in stable transmission and
`the highest titers (Fig. 4, lanes A, E, L, and N).
`The LCR was first functionally defined in vivo in transgenic
`mice by using a 20-kb LCR fragment (10, 11), a 6.5-kb
`fragment (15), or single HS sites (17-22, 27-29). Studies in
`animals suggested that the LCR confers position-independent
`expression, based on the calculation that ,B-globin expression
`per transgene copy is relatively constant (7). However, it
`should be noted that transgenic studies involving either com-
`plete LCR or core sequences generally relied on the analysis
`of cells possessing more than one copy per cell. In contrast, our
`use of retrovirus-mediated gene transfer made it possible to
`examine the issue of position independence under conditions
`
`Downloaded from https://www.pnas.org by 163.116.135.92 on January 21, 2023 from IP address 163.116.135.92.
`
`SKI Exhibit 2024
`Page 4 of 5
`
`

`

`6732
`
`Genetics: Sadelain et aL
`
`Proc. Natl. Acad. Sci. USA 92 (1995)
`
`where one copy of the transcription unit per cell is integrated
`in a precise reproducible way and in the absence of any
`selection (50). In a panel of MEL cell clones representing
`independent integration sites, we found that expression was
`detectable in all clones, yet ranged from 4% to 146% (Table
`1) of human ,B-globin expression in the Hull cell line in
`single-copy clones (the variability decreases, as expected, in
`multicopy clones). Our data therefore suggest that the LCR, in
`this configuration, acts more like a classical enhancer than the
`proposed function of an LCR (reviewed in refs. 7-10 and 51).
`The absence of strictly position-independent expression (10,
`14, 15) we observe raises the issues of whether the LCR can
`truly confer position-independent expression when present at
`one copy per cell, whether the compact arrangement of the
`transcription and chromatin regulators within our vector leads
`to suboptimal interactions and inefficient LCR activity, or
`whether it is specifically the combination of core elements
`which we have employed which lack this ability (see refs. 17-22
`for HS2, 27 and 28 for HS3, and 29 for HS4). Whatever the
`explanation, our data strongly suggest that it is important to
`carefully reevaluate the characteristics of larger LCR-
`containing sequences in the context of single-copy insertions.
`For the purposes of gene therapy the most critical test of the
`LCR sequences will be the introduction of the virus genomes
`described above into hematopoietic cells followed by their
`transplantation. The inability to obtain chromosome position-
`independent expression of the human ,B-globin gene in the
`context of single-copy insertions would make the prospects for
`effective genetic treatment of hemoglobinopathies less likely
`than previously believed.
`
`M.S. thanks Drs. I. Riviere and M. Goodell for reviewing the
`manuscript and the Medical Research Council of Canada for support.
`M.S. is a recipient of the Medical Research Council Centennial
`Fellowship. C.H.J.W. was supported by the Undergraduate Research
`Opportunities Program at the Massachusetts Institute of Technology.
`This work was supported by National Institutes of Health Grant
`HL37569 to R.C.M.
`
`1.
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`8.
`
`9.
`10.
`
`11.
`
`12.
`
`13.
`
`Dzierzak, E. A., Papayannopoulou, T. & Mulligan, R. C. (1988)
`Nature (London) 331, 35-41.
`Karlsson, S., Bodine, D. M., Perry, L., Papayannopoulou, T. &
`Nienhuis, A. W. (1988) Proc. Natl. Acad. Sci. USA 85, 6062-6066.
`Bodine, D. M., Karlsson, S. & Nienhuis, A. W. (1989) Proc. Natl.
`Acad. Sci. USA 86, 8897-8901.
`Bender, M. A., Gelinas, R. E. & Miller, A. D. (1989) Mo. Cell.
`Bio. 9, 1426-1434.
`Tuan, D., Solomon, W., Li, Q. & London, I. M. (1985) Proc. Natl.
`Acad. Sci. USA 82, 6384-6388.
`Forrester, W. C., Takegawa, S., Papayannopoulou, T., Stamatoy-
`annopoulos, G. & Groudine, M. (1987) Nucleic Acids Res. 15,
`10159-10177.
`Dillon, N. & Grosveld, F. (1993) Trends Genet. 9, 134-137.
`Crossley, M. & Orkin, S. H. (1993) Curr. Opin. Genet. Dev. 3,
`232-237.
`Engel, J. D. (1993) Trends Genet. 9, 304-309.
`Grosveld, F., van Assendelft, G. B., Greaves, D. R. & Kollias, G.
`(1987) Cell 51, 975-985.
`Ryan, T. M., Behringer, R. R., Martin, N. C., Townes, T. M.,
`Palmiter, R. D. & Brinster, R. L. (1989) Genes Dev. 3, 314-323.
`Peterson, K. R., Clegg, C. H., Huxley, C., Josephson, B. M.,
`Haugen, H. S., Furukawa, T. & Stamatoyannopoulos, G. (1993)
`Proc. Natl. Acad. Sci. USA 90, 7593-7597.
`Gaensler, K. M. L., Kitamura, M. & Kan, Y. W. (1993) Proc.
`Natl. Acad. Sci. USA 90, 11381-11385.
`
`14.
`
`15.
`
`16.
`
`17.
`
`18.
`
`19.
`
`20.
`
`21.
`
`22.
`
`23.
`
`24.
`
`25.
`
`26.
`27.
`
`28.
`
`29.
`
`30.
`
`31.
`
`32.
`
`33.
`
`34.
`
`35.
`36.
`
`37.
`38.
`39.
`
`40.
`41.
`42.
`
`43.
`
`44.
`45.
`
`46.
`
`47.
`
`48.
`
`49.
`
`50.
`
`51.
`
`van Assendelft, G. B., Hanscombe, O., Grosveld, F. & Greaves,
`D. R. (1989) Cell 56, 969-977.
`Talbot, D., Collis, P., Antoniou, M., Vidal, M., Grosveld, F. &
`Greaves, D. R. (1989) Nature (London) 338, 352-355.
`Fraser, P., Hurst, J., Collis, P. & Grosveld, F. (1990) NucleicAcids
`Res. 18, 3503-3508.
`Ryan, T. M., Behringer, R. R., Townes, T. M., Palmiter, R. D. &
`Brinster, R. L. (1989) Proc. Natl. Acad. Sci. USA 86, 37-41.
`Curtin, P. T., Liu, D., Liu, W., Chang, J. C. & Kan, Y. W. (1989)
`Proc. Natl. Acad. Sci. USA 86, 7082-7086.
`Talbot, D., Philipsen, S., Fraser, P. & Grosveld, F. (1990) EMBO
`J. 9, 2169-2178.
`Caterina, J. J., Ryan, T. M., Pawlik, K. M., Palmiter, R. D.,
`Brinster, R. L., Behringer, R. R. & Townes, T. M. (1991) Proc.
`Natl. Acad. Sci. USA 88, 1626-1630.
`Liu, D. P., Chang, J. C., Moi, P., Liu, W., Kan, Y. W. & Curtin,
`P. T. (1992) Proc. Natl. Acad. Sci. USA 89, 3899-3903.
`Caterina, J. J., Ciavatta, D. J., Donze, D., Behringer, R. R. &
`Townes, T. M. (1994) Nucleic Acids Res. 22, 1006-1011.
`Forrester, W. C., Novak, U., Gelinas, R. & Groudine, M. (1989)
`Proc. Natl. Acad. Sci. USA 86, 5439-5443.
`Collis, P., Antoniou, M. & Grosveld, F. (1990) EMBO J. 9,
`233-240.
`Ney, P. A., Sorrentino, B. P., McDonagh, K. T. & Nienhuis,
`A. W. (1990) Genes Dev. 4, 993-1006.
`Talbot, D. & Grosveld, F. (1991) EMBO J. 10, 1391-1398.
`Philipsen, S., Talbot, D., Fraser, P. & Grosveld, F. (1990) EMBO
`J. 9, 2159-2167.
`Philipsen, S., Pruzina, S. & Grosveld, F. (1993) EMBO J. 12,
`1077-1085.
`Pruzina, S., Hanscombe, O., Whyatt, B., Grosveld, F. & Phil-
`ipsen, S. (1991) Nucleic Acids Res. 19, 1413-1419.
`Novak, U., Harris, E. A. S., Forrester, W., Groudine, M. &
`Gelinas, R. (1990) Proc. Natl. Acad. Sci. USA 87, 3386-3390.
`Plavec, I., Papayannopoulou, T., Maury, C. & Meyer, F. (1993)
`Blood 81, 1384-1392.
`Chang, J. C., Liu, D. & Kan, Y. W. (1992) Proc. Natl. Acad. Sci.
`USA 89, 3107-3110.
`Guild, B. C., Finer, M. H., Housman, D. H. & Mulliga