c¿-~ 2'l'2JOURNAL OF SURGICAL RESEARCH 56, 510-517 (1994)

Liver-Directed Gene Therapy: Evaluation of Liver SpecificPromoter Elements

DANIEL G. HAFENRICHTER, M.D.,* KATHERINE P. PONDER, M.D., t:!: STEVEN D. RETTINGER, M.D.,*SUSAN C. KENNEDY, B.S.,:!: XIAOYAN Wu, M.S.,:!: ROBERT So. SAYLORS, M.D.,:!:

ANDM. WAYNEFLYE, M.D., PH.D.*Departments ol.Surgery, t lnternal Medicine, and :l:Biochemistry & Molecular Biophysics, Washington University School 01 Medicine,

Sto Louis, Missouri 63110

Presented at the Annual Meeting of the Association for Academic Surgery, Hershey, Pennsylvania, November 10-13, 1993

Liver-directed gene therapy could dramatically alterthe therapy of many inherited hematologic and meta-bolic diseases. We have developed a rapid, reliable,low-mortality method for the in vivo delivery in rats ofretroviral vectors 24 hr after 70% hepatectomy by in-

traportal injection during hepatic in-flow occlusion(HIFO). Using the human cxl-antitrypsin (hAAT) re-porter gene, we found in vivo that up to 100;0 ofhepato-cytes integrated the provirus, and serum hAA T proteinlevels were sustained for up to l year. Despite high invivo transduction efficiencies, gene expression at themRNA level is disappointingly low compared to in vi-tro transduced NIH 3T3 or Hepa Al tissue culture cells.In this report, LNL-6-derived retroviral vectors (RV)were combined with ORe of two strong, liver-specificpromoters, murine albumin or human cxl-antitrypsin,an upstream insertion of a trimer of hepatocyte nuclearfactor-3 (HNF-3) binding sites, and the hAAT reportergene. HNF-3 has been demonstrated to increase in vi-tro transcriptional activity [23]. Twenty-four hoursafter 700;0 hepatectomy, lO-raId concentrated (bymethotrexate-resistant titer) RV -producing cell super-natant was given intraportally during a 3-min HIFO.Serum hAA T le veIs as quantitated with ahuman spe-cific ELISA were sustained for ayer 40 weeks with allof the liver-specific promoter constructions. However,the hAA T protein production with the murine albuminpromoter retroviral constructs decreased with time,but was sustained at levels approximately 800;0 of theinitial serum peak levels with the constructs containingthe hAA T promoter. We conclude that the retroviralvectors containing liver-specific promoters can lead tosustained, transduced gene express ion and the quanti-tative evaluation of in vivo promoter strengths is es-sential for the development of high expressing retro-viral vectors necessary for hepatic gene therapy inhumans. @ 1994 Academic Press. loc.

against infection. Although these deticiencies can betreated by repeated infusions of the particular protein,this is expensive, carries an associated risk ofviral infec-"tion, and only temporarily ameliorates the disease mani-festations. Liver transplantation has been used tocorrect serum protein deticiencies [1] but is limited bydonar organ availability and requires immunosuppres-sion. The transfer of a functional gene into the cells of agenetically deticient individual should correct the clini-cal manifestations of the disease by providing long-termproduction of the deticient gene producto Accordingly,the liver has been a primary target organ for somaticgene therapy. Both ex vivo [2, 3,16,41] and in vivo [4-12]methods have been used to transfer genes into mamma-lían hepatocytes.

The ex vivo approach requires hepatocyte harvest, invitro transduction with retrovirus, and reintroduction ofthe transduced hepatocytes into the portal circulation.While Chowdhury et al. [2] and Kay et al. [3] haveachieved reasonable expression of low-density lipopro-tein receptor (LDL-R) and human al-antitrypsin(hAA T) genes, respectively, limitations to ex vivo genetherapy include the potential for contamination duringlarge-scale hepatocyte culture (108 hepatocytes/kg recip-ient body wt), limitation of the number of hepatocyteswhich can be reintroduced (1-2%) due to the potentialfor portal venous obstruction, and the requirement of apartial hepatectomy to harvest hepatocytes.

In vivo delivery offoreign genes to hepatocytesin situhas been accomplished by several approaches. The intra-venous injection of liposome-coated [4] or asialoglyco-protein-coated [5] plasmid DNA or adenoviral vectors[45] has resulted in gene expression in vivo, whichusually declines over time due to the instability of theextrachromosomal DNA. Retroviral vectors have beendelivered to the liver in vivo by intraparenchymal [6] orintraportal injection [7, 8], asanguineous liver perfusion[9, 10], and hepatic inflow occlusion/portal vein injec-.[11,

INTRODUGrION

The liver synthesizes many serum proteins which playpivotal roles in hemostasis, metabolism, and protection ,12], hepatic ischemia [10], andcarbon tet-

5100022-4804/94 $5.00Copyright @ 1994 by Academic Press, [nc.All rightB of reproduction in any form reBerved.

HAFENRICHTER ET AL.: LIVER-SPECIFIC PROMOTER ELEMENTS 511

rachloride injury [8] have been used to induce hepato-cyte replication.

In attempts to increase gene expression, retroviralvectors containing strong viTal promoters have beenused. While these promoter constructs are extremely ac-tive in tissue culture cells, they are poorly expressed inhepatocytes and in vivo attenuate with time [2, 3, 7, 12-14, 45]. Therefore, the use of cellular promoters has beenproposed as a means of providing sustained expression[2, 7, 11]. While insertion of the promoter for the largesubunitofmurine RNApolymerase II (Pol-lI) into retro-viral vectors containing the hAA T reporter gene sus-tained expression for ayer 1 year, decreases in seruinhAA T levels algo occurred. This was felt to likely be dueto the loss of function from the L TR and not from theinternally placed Pol-1I promoter [11].

Therefore, in this study, the expression of retroviralvectors containing the strong, liver-specific 810-bp mu-rine albumin (mAlb) and the 347-bp hAAT promotersplus the trimers ofthe DNA binding sequence for hepa-tocyte nuclear factor-3 (HNF-3) were measured in vivoby serum hAAT production. In vitro, HNF-3 increasedthe hAAT transcription from the mAlb [23], while invivo, its presence appeared to exert a stabilizing inftu-ence on long-term serum hAAT protein expression.

albumin promoter was obtained from pAT2 (from Dr.Kenneth Zaret, Brown University, Providence, RI) byHpalI digestion (which cuts just downstream from thetranscription start Bite) and blunt end formation withT4 polymerase. The BamHI linkers were ligated, di-gested with BamHI and EcoRI, and ligated into theEcoRIj BamHI Bite of SP72. A BglIIj BamHI fragment ofmAlb-SP72 containing the 810-bp albumin promoterwas ligated into the unique BgllI Bite of hAA T -ÁL TR tocreate mAlb-hAAT-ÁLTR.

Construction of hAAT-hAAT-ÁLTR (Fig. 2). The347-bp fragment of the hAAT promoter from the BglIlBite at -347 to the SmaI Bite at +10ofthe hAATgene 3Ewas BclI linker ligated. After BclI and BgllI digest(which cuts at -347), the DNA was cloned into the BglIlBite of hAAT-ÁLTR, regenerating the unique 5' BgllIBite, and creating hAAT-hAAT-ÁLTR.

lnsertion of HNF-3 binding sites (Figo 2)0 A 200-bpfragment containing HNF -3 binding Bite trimers wereobtained from Dr. Kenneth Zaret and was digestedwithEcoRI and cloned into the EcoRI Bite of pSP72. TheHNF -3 fragment was removed by digesting with BamHIand BgllI and placed into the mAlb and hAA T promotercontaining retroviral vectors via the unique 5' BgllI Bite.

Creation of retroviral producer celllineso The ampho-tropic GP+envAMI2 [26] murine fibroblast line wasmaintained in Dulbecco's modified Eagle's medium sup-plemented with 10% heat-inactivated calf serum (Hy-clone Laboratories, Logan, UT), penicillin (105 Ujliter),streptomycin (105 mgjliter) (DMEM). After calciumphosphate transfection with retroviral vector DNA, theGP+envAMI2 cells were cultured for 14 days in 250 nMmethotrexate (Sigma Chemical, Sto Louis, MO) supple-mented DMEM. Supernatant from 50 methotrexate-re-sistant (MtxR) GP+envAMI2 clones were screened forthe production of retrovirus by infecting NIH 3T3 cellsand measuring hAA T production 72 hr later. MtxR titerswere determined on NIH 3T3 cells [27], and high-titerclones were proveo free of replication competent helpervirus by amplification and marker rescue assay [27].

In vivo hepatocyte transduction protocolo Age-matched adult male Sprague-Dawley rats (SASCO,amaba, NE) weighing 200-275 g, received rodent chowand tap water ad libitum. Standard NIH care with a 12-hlight-dark cycle was provided. Twenty-four hours after70% hepatectomy, in vivo hepatocyte transduction was~

MATERIALS ANO METHOOS

hAA T ELISA. hAA T was measured by sandwichELISA [28] using the IgG fraction of goat anti-hAAT(Atlantic Antibodies, Scarborough, NJ; Lot No. 60206)as the first and the second antibody. The second anti-

Construction o{ retroviral vectors. The Moloney mu-rine leukemia virus (Mo-MLV)-based LNL-6 retroviralvector (LTR-hAAT) contains the LTR promoter, thehAAT cDNA reporter gene, the encephalomyocarditisvirus internal ribosome entry site (IRES), and the mu-tant dihydrofolate reductase gene and was cloned as pre-viously described [11]. In arder to quantitatively evalu-ate the mAlb and hAA T liver-specific promoters, a retro-viral vector was constructed which contained a 178-bpdeletion in the enhancer region ofthe 3' LTR (Fig. lA).Transfection of packaging cells leads to the integratedvector in the packaging cells (Fig. lB), which transcribesmRN A from the L TR and the internal promoters(arrows). Injection of preparations of the retroviral vec-tor produced by the packaging cells into rats leads to theintegrated provirus within the target cell genome (Fig.lB), which predominantly transcribes mRNA from theinternal promoter (arrow). Thus, as a result of the 3'enhancer deletion, the majority of the transcripts willoriginate from the internal promoter, thus facilitatingquantitative comparisons of liver specific promoter ele-ments. Therefore, the LTR-hAAT-ALTR retroviralvector (RS 32392- hAA T) which contains a 178-bp dele-tion in the 3' L TR enhancer region was created by re-moval ofthe sequences from the 5' PvulI site tothe XbaIofLTR-hAAT sites as described [25]. This deletion willsubsequently be referred to as the AL TR deletion.

Construction o{ mAlb-hAAT-ALTR (Fig. 2). Tocreate mAlb-SP72, the 810-bp fragment of the mouse

512 JOURNAL OF SURGICAL RESEARCH: VOL. 56, NO. 6, JUNE 1994

A. LTREnhancerDlrect Repeats CAAT TATA pA

~ ~ ~I -I -I

72 bp 72 bp

Mo-MLV L TR

I"'0/1

~

tx...,

IDeletion (178 bp)

PackagingCellRetroviralGenome

del enh

""'::;:7del enh

"":::::7

p h I I , d

IntegratedProvirusWithinTransducedCel!

FIG. l. (A) The structure of the intact Moloney murine leukemia virus (Mo-MLV) I~TR is shown with the relutive positions of thcelements depicted: the 72-bp direct repeats ofthe enhancer, the polyadenylation signul sequence (pA), and the CAAT and the TATA transcrij>-tional elements. A deletion which encompasses portions of the two tandem repeat regions of the L TR enhancer wus created by restrictionenzyme digestion, as described under Materials and Methods. (B) The retroviral genome which is present in the packaging cellline, containingthe 3' L TR deletion (shown by del enh and AL TR), which is transferred to the 5' L TR after transduction of the target cell. The promoter (p),hAAT reporter gene (h), IRES sequences (1), and mutant dihydrofolate reductase gene (d) are abbreviated for clarity. In the packaging celllinegenome, transcription can initiate from the L TR or the internal promoter. As a result of the L TR enhancer deletion, transcription from theintegrated provirus genome predominantly initiates from the internal promoter [25, 35].

body was labeled with hors~radish peroxidase (Conju-Quick; Midwest Scientific, Sto Louis, MO), and devel-oped with 3,3',5,5'-tetramethylbenzidine dichloride sub-strate (Sigma). The lower sensitivity limit of the assaywas <3 ngfml hAAT.

duced at a low multiplicity of infection. After se le ctio nwith methotrexate, resistant cells should have a singlecopy of the retrovirus integrated, which facilitates in vi-tro comparison of promoters. Pools of transduced MtxRHepa Al and NIH 3T3 cells were analyzed for their pro-duction of hAAT protein (Table 1). For the mAlb pro-motel constructs, the insertion of the HNF -3 bindingsites produced threefold lower amounts of hAA T proteinon the NIH 3T3 fibroblasts, which are devoid of liver-specific transcription factors, but no change in hAA Tprotein production was measured in the hepatoma cells.The hAAT-hAAT-ALTR and HNF-3-hAAT-hAAT-AL TR promoter constructs demonstrated similar pro-duction of hAA T protein by the NIH 3T3 cells in vitro(Table 1). We could not quantitate the hAAT produc-tion by the mAlb-hAAT-ALTR-transduced Hepa Al

RESULTS

In vitro evaluation of liver-specific promoter construc-tions. Retroviral constructs were transfected into am-photropic GP+envAM-l2 packaging cells and selectedfor Mtx resistance. With retroviral supernatants fromhighest titer GP+envAM-l2 clone producing packagingcelllines for each of the four promoter constructs, NIH3T3 fibroblasts and Hepa Al hepatoma cells were trans-

HAFENRICHTER ET AL.: LIVER-SPECIFIC PROMOTER ELEMENTS 513

Sgl/l

HNF -3 -mAlb

89//1

hAAT

HNF-3-hAAT8¡¡/1/

hAAT *DHFR

'1'+

FIG.2. Description of retroviral vectors containing the liver-specific promoter constructions. The LNL-6-derived retroviral vector back-bone (RS3891) [111, with a 3' LTR enhancer deletion, is abbreviated as ~LTR. The murine albumin promoter was inserted into the BglIIrestriction site to crea te Alb-hAA 'I'-~L TI{ (abbreviated here as mAlb), and a trimer of hepatocyte nuclear factor-3 binding sites was inserted into tlle regenerated lJgll1 restriction site upstream from the mAlo promoter, creating HNF-3-mAlo-hAAT.~LTR (aboreviated as HNF-3-mAlo). 'l'he hAA'I' promoter containing vectors hAAT.hAA 'I'-~L 'l'R and HNF-3.hAAT-hAAT-~L TR (abbreviated hAA'I' and HNF -3-hAA T.respectively) were created in a similar fashion, and are descrioed under Materials and Methods.

cells due to insullicient titer and the low infectability ofthe Hepa Al cells.

Amphotropic packaging cell clones had MtxR retro-viral titers determined on NIH 3T3 and Hepa Al cells asestimated by limiting dilution analysis of unconcen-

TABLEl

In Vitro Transduced Gene Expression of the Retro-vira} Vectors Containing Liver-Specific Promoters

hAAT protein production(ng/24 hr/lO6 cells)

NIH 3T3 cells Hepa Al cellsPromoter construction

11337.5

447459

**

203217

mAlbHNF-3-hAATHNF-3-

trated supernatant [11] for the mAlb-hAAT-ÁLTR- andthe hAAT-hAAT-ÁLTR-containing constructs rangingfrom 10,000 to 60,000 colony forming units (cfu)/ml.

Long-term expression 01 liver-specific promoters alterin vivo transduction. One of the principal goals of genetherapy is long-term expression of the transduced gene.Rats that were transduced via the intraportal injectionof 24-fold concentrated (by volume and 10-fold by MtxRtiter) packaging cell supernatant containing the retro-viral construct mAlb-hAAT-ÁLTR during hepatic in-flow occlusion, initially expressed high levels of serumhAA T in two of three rats, but by 40 weeks post-trans-duction, expression had decreased to 48% of post-trans-duction Week 1 values (Fig. 3). The insertion of theHNF -3 binding sites into the mAlb promoter constructresulted in lower early levels of hAA T protein in theserum, but expression and degree of attenuation (46%)comparable to those of the mAlb-hAAT-ÁLTR vectortransductions ayer the 40-week study periodo

With the hAAT promoter, seven of seven rats ex-pressed the hAA T gene for ayer 16 weeks. Attenuationof serum hAA T protein levels was less than that with themAlb constructs, since the hAAT-hAAT-ÁLTR-trans-duced rats produced 80% oftheir Week 1 post-transduc-tion levels at 16 weeks post-transduction. Rats trans-duced with the HNF-3-hAAT-hAAT-ÁLTR construct

Note. NIH 3T3 (ATCC No. 1658) [46J fibroblasts and Hepa Al [47Jhepatoma cells were transduced at a low multiplicity of infection withGP+enuAM-12 packaging cell supernatants from each of the fourliver-specific promoter constructs. The transduced NIH 3T3 andHepa Al cells underwent methotrexate selection, and the hAAT pro-duction (ng/24 hr/106 cells) was quantitated by ELISA. Asterisks (.)denote nondetectable hAA T production due to insufficient numbersof transduced Hepa Al cells after MTX selection (see Results).

mAlb

hAAT

514 JOURNAL OF SURGICAL RESEARCH: VOL. 56, NO. 6, JUNE 1994

::;E-C'IC~

S.cE~...GItn

gene expression with viral promoters include D NA meth-ylation [8, 17], negative regulatory proteins [18, 19, 20,21], and the absence oftranscription factors in quiescenthepatocytes [15, 22].

The use of cellular promoters has been proposed as ameans of solving the viral promoter attenuation prob-lem. Chowdhury et al. [2] utilized a chicken fJ-actin pro-motel to direct the gene for LDL-R and observed noattenuation in vivo after 3 months, although levels of invivo transduction were not quantified. We have used ourretroviral transdaction method [11, 12] to compare theability of retroviral vectors containing the native L TRretroviral promoter or an inserted cellular Pol-1l pro-motel to express the hAA T at the mRNA and serumprotein levels [11]. While reasonable transduction fre-quencies of 10-15% were obtained, RNA levels weremuch lower than those for transduced tissue culturecells, but sustained serum protein expression wasachieved for over 1 year. This suggests that in vivo tran-scription may be inappropriately low from the construct.

The efficacy of liver-specific promoters to drive high-level, sustained reporter gene expression has been dem-onstrated in cultured hepatoma celllines, hepatocytes,and in transgenic mouse systems [14, 41, 44]. Kay et al.[7] inserted the liver-specific murine albumin promoterfenhancer construction into a retroviral vector with in-tact L TR sequences which directed the expression of thehAA T reporter gene. Serum hAA T protein levels aver-aged 500 ngfml and gene expression was sustained Cor atleast 6 months, suggesting that this promoter was 5- to10-fold stronger than the Pol-11 or Mo-MLV L1'R pro-moters. This promoter construct had a 2.0-kb deletion(M. Kay, personal communication). In contrast, trans-duction with the liver-specific phosphoenolpyruvatecarboxykinase promoter containing retroviral vectors

o 10 20 30 40

Weeks Post Transductlon

FIG.3. Long-term in vivo gene expression ofthe rats transducedwith the retroviral vectors containing the liver-specific murine albu-min promoter constructs.

maintained 88% of their Week 1 post-transduction val-ues at 40 weeks (Fig. 4).

DISCUSSION

The liver is an excellent target organ for in situ retro-viral gene therapy. Portal venous and hepatic arterialblood bathes hepatocytes through large fenestrations inthe sinusoidal endothelium, permitting direct contact ofhepatocytes with large particles such as retroviruses. AI-though retroviral infection of the resting liver is ineffi-cient due to the requirement of cell division for the geno-mic integration of retroviruses, partial hepatectomy re-sults in a coordinated burst of hepatocyte cell division,with up to 44% of hepatocytes proliferating at 25 hrposthepatectomy [36]. The subsequent long lifespan ofquiescent hepatocytes should facilitate stable and long-term express ion of a transduced gene. In addition, manyof the proteins deficient in inherited metabolic and he-matologic diseases are synthesized within the liver,where they undergo post-translational modificationsnecessary for full functional activity.

Although a number of investigators have used retro-viral vectors to transduce up to 0.5-10% of hepatocytesof animals, the inability to achieve high levels of expres-sion remains the single major obstacle to successful he-patic gene therapy. In attempts to increase gene expres-sion, retroviral vectors containing strong viral pro-moters have been used. In general, viral promoters,which are extremely active in tissue culture cells, areexpressed poorly by retroviral vectors in vivo. For exam-pIe, the cytomegalovirus promoter in vitro demonstrateshigh levels of gene expression in hepatocytes [3,7, 12, 13,14], but in vivo gene expression rapidly attenuates [3, 7].Possible mechanisms responsible for the decrease in

o 10 20 30 40

Weeks Post Transduction

FIG.4. Long-term in vivo gene expression ofthe rata transducedwith the retroviral vectors containing the liver-specific human cx,-an-titrypsin promoter constructs.

HAFENRICHTER ET AL.: LIVER.SPECIFIC PROMOTER ELEMENTS 515

long-term gene expression since much of the LTR tran-scription is inactivated [25]. The in vivogerie expressionobtained with our mAlb-hAAT-ALTR retroviral vectoris consistent with that of Kay et al. [7] using the sameL TR enhancer deleted retroviral vector backbone, themurine albumin promoterfenhancer, and a lower num-ber of retroviral particles injected into hepatectomizedmice. We observed that for the mAlb-hAAT-ALTR vec-tor, the serum hAA T protein expression attenuated to48% of Week 1 values over the 10-month study period,which is consistent with the decrease in hAAT gene ex-pression over time observed by Kay et al. [7] with theliver-specific albumin promoter f enhancer. However,the hAAT-hAAT-ALTR vector-transduced rats in ourstudy demonstrated a minimal (12-20%) decrease overthe 40-week study periodo

In the liver, multiple cis-acting transcription factorscontribute to the high level of gene expression in thehepatocyte compared to other cells [43]. Among thetranscription factors is the HNF -3 binding activity thatconsists of three distinct proteins (HJ'lF-3a, HNF-3(J,and HNF -3'Y) each of which have been cloned and shownto be preferentially expressed in the liver and to a lesserextent in a subset of other tissues [42]. Oligomers ofHNF-3 binding sites placed upstream of a murine albu-min promoter containing vector transfected into hepa-toma cells resulted in a 30-fold increase in transcrip-tional activity compared to the murine albumin pro-moter alone [23]. In vivo, the addition of the HNF-3binding sites to the mAlb-hAA T -AL TR had no apparenteffect on the stability of long-term hAA T protein produc-tion. However, for the hAA T promoter constructs, theHNF-3-hAAT-hAAT-ALTR vector resulted in a lower(12%) decrease of in vivo gene expression over the basalhAAT-hAAT-ALTR promoter (20%) over the 40-weekstudy periodo These results indicate that the hAAT-hAAT-ALTR construct is more stable in vivo over timethan the mAlb promoter construct, an effect which maybe supplemented by the insertion of the HNF -3 bindingsites. This may be a result of interactions between thetranscription factor binding Bite within each liver-spe-cific promoter construction [23, 42, 43] or the upstreamALTR sequence [18, 19,20,22,25].

Several groups have reported low titers using retro-viral vectors with deletions in the L TR transcriptionalunit [7, 31, 32]. Our titers were substantially lower thanthose of Soriano et al. [25] and Kay et al. [7] using theAL TR deletion in a different retroviral backbone and aninternally placed murine albumin promoterfenhancer.Since only full-length retroviral RNAs which contain anintact psi packaging signal, 5' and 3' L TR regions can beeffectively packaged into a retroviral particle andsubse-quently integrate into a target cell chromosome, the initi-ation of transcripts from the internal promoter may re-sult in lower retroviral titers. Although the retroviraltiters of the liver-specific promoter constructions con-taining the ALTR deletion was 40- to 100-fold lower

led to extremely low levels of human growth hormonereporter gene expression [6]. Therefore, we employed ret-roviral vectors containing either the strong, liver-spe-cific murine albumin or human cxl-antitrypsin pro-moters to increase and stabilize hAA T gene expressionin the transduced rats.

Attempts to utilize tissue-specific regulatory se-quences in retroviral vectors may be complicated by thepresence of powerful regulatory elements in the L TRregions which may interfere with the functions of thetissue specific signals [32]. To limit this problem ofvaria-tion in gene expression by affected internal promoters,Yu et al. [31] used a self-inactivating (SIN) vector inwhich the enhancer and the "CAA T box" were deletedfrom the L TRs. Yee et al. [32] employed a disabled retro-viral vector in which the enhancer, the CAAT box, andthe transcriptional start site were deleted from the L TRregions. Increased transcription derived from the inter-nal promoters was demonstrated with each of these con-structions.

A Mo-MLV-based retroviral vector with an internalpromoter (Pol-lI) inserted can initiate transcriptionfrom either the 5' LTR or the internal Pol-lI promoter[11]. To diminish the possible influence of viral tran-scription on the quantitation of gene expression fromthis internally placed promoter, we employed the ALTRMo-MLV-based retroviral vector which creates a dele-tion of the 5' LTR enhancer upon integration into thetarget cell genome [25]. The 178-bp Puull/XbaI frag-ment (Fig. 1) which was removed, encompasses most ofthe two 75-bp tandem repeats comprising the LTR en-hancer [32]. The U3 regio n in the 3' L TR of a infectiousretroviral RNA serves as a template for the formation ofboth U3 regions during replication of retroviruses [34].Therefore, a deletion in the U3 enhancer region of the 3'L TR will be transferred to both L TRs in the provirus.This reduces the transcriptional activity of the L TRpromoter by as much as 85-90% [25,35], but exerts min-imal effects upon an internally placed promoter [33]. Inaddition, the deletion does not appear to effect the stabil-ity of proviral expression, as this vector backbone hasbeen used by Kay et al. [7] and Chowdhury et al. [2] todrive gene expression controlled by internally placedpromoters for at least 3 and 6 months, respectively.

An advantage of using retroviral vectors is sustainedgene expression resulting from the highly efficient retro-viral integration into the target cell chromosome. Whenwe followed the serum hAA T expression in rats trans-duced with the Pol-lI vector long term, the serum hAA Tprotein le veis decreased to steady-state le veis after 12weeks, which were an average of 31 % of init~al Week 1values [11]. Mapping ofthe transcription initiation sitein these rats suggested that much ofthe observed attenu-ation of hAA T gene expression was due to a decrease inthe L TR-initiated transcript [11]. The insertion of liver-specific promoters into retroviral vectors with deletedL TR enhancer regions facilitates the assessment ofthe

516 JOURNAL OF SURGICAL RESEARCH: VOL. 56, NO. 6, JUNE 1994

than those of our Pol-II-containing vectors (11), theserum hAAT protein production was only 2- to 7-foldlower in the mAlb and hAA T promoter constructions.This suggests that the liver-specific promoter construc-tions are between 10- and 14-fold stronger than the con-stitutive Pol-II promoter.

We conclude that the effect of internally insertedliver-specific promoters and promoter elements in vivois highly dependant upon the individual promoter andretroviral vector utilized. These results demonstrate theimportance of in vivo comparisons of promoter- andgene-regulatory elements. We are currently analyzingthe liver-specific promoter constructs at the proviralDNA level to quantitatively compare the promoter ele-ment strengths. Our retroviral transduction method,which allows the rapid assessment of gene regulatoryelements and their interactions in vivo, should facilitatethe development of potent retroviral vectors for thetreatment of human disease.

ACKNOWLEDGMENT

The authors acknowledge the expert and dedicated secretarial assislance of Mrs. Theresa Belgeri in the preparation of the manuscript.

REFERENCES

11. Rettinger, S. D., Kennedy, S. C., Wu, X., Saylors, R. L., Ha-fenrichter, D. G., Flye, M. W., and Ponder, K. P. Liver directedgene therapy: Quantitative evaluation of promoter elements us-ing in vivo retroviral transduction. Frac. Natl. Acad. Sci. USA91: 1460, 1994.

12. Rettinger, S. D., Ponder, K. P., Saylors, R. L., Kennedy, S. C.,Hafenrichter, D. G., and Flye, M. W. In vivo hepatocyte trans-duction with retrovirus during in-ftow occlusion. J. Surco Res.54: 418, 1993.

13. Boshart, M., Weber, F., Jahn, G., Dorsch-Hasler, K., Flecken-stein, B., and Schaffner, W. A very strong enhancer is locatedupstream of an immediate early gene of human cytomegalovirus.Cel/ 41: 521, 1985.

14. Ponder, K. P., Dunbar, R. P., Wilson, D. R., Darlington, G. J.,and Woo, S. L. C. Evaluation of relative promoter strength inprimary hepatocytes using optimized lipofection. Hum. GeneTher. 2: 41, 1991.

15. Kay, M. A., Rothenberg, S., Landen, C. N., Bellinger, D. A., Le-land, F., Toman, C., Finegold, M., Thompson, A. R., Read, M. S.,Brinkhous, K. M., and Woo, S. L. C. In vivo gene therapy ofhemophilia B sustained partial correction in factor IX-deficientdogs. Science 262: 117, 1993. "

16. Wilson, J. M., Chowdhury, N. R., Grossman. M., Wajsman, R.,Epstein, A., Mulligan, R. C., and Chowdhury, J. R. Temporaryamelioration of hyperlipidemia in low density lipoprotein-defi-cient rabbits transplanted with genetically modified hepatocytes.Frac. Natl. Acad. Sci. USA 87: 8437, 1990.

17. Jaenisch, R., Schnieke, A., and Harbers, K. Trealment of micowilh 5-azacylidine eflicienlly aclivlIles silenl relrovirlll genomesin difl'erenllissues. l'rllc. Nall. Acad. Sci. U.~A 82: 1451,1985.

18. TsukiYllma, T., Niwa, O.,and Yokoro, K. Mechal1ism ofsuppres-sion of lhe long terminal repeal of Moloney leukemia virus inmouse emhryonal carcinoma cells. MIli. Cel/. Birll. 9: 4G70. 1989.

19. Gorman, C. M., RiKhy, P. W. .1., and (.une, O. NeKalive relt\lla-lion ofviral enhancer!l in undifl'erenlinled emhryonic stem cell!l.Cel/ 42: 519, 1985.

20. Grez, M., ZorniK, M., Nowock, .1.,alld ZieKler, M. A sinKle pointmulalion aclivates the Moloney murine leukemin virus long ter-minal repeat in emhryonalst.em cells. J. Virol.65: 4691,1991.

21. FlannKan, J. R., Becker, K. G., Ennist, D. L., Gleason, S. L.,Driggers, P. H., Levi, B.Z., Appella, E., and Ozato, K. Cloning ofa negalive transcription factor that binds to the upstrenm con-served region of Moloney murine leukemia virus. Mol. Cel/. Biol.12: 38, 1992.

22. Speck, N. A., and Ballimore, D. Six nuclear factors interact withthe 75 base pair repeat of the Moloney murine leukemia virusenhancer. Mol. Cel/. Biol. 7: 1101, 1987.

23. DiPersio, C. M., Jackson, D. A., and Zaret, K. The extracellularmatrix coordinately modulates liver transcriplion factors andhepatocyte morphology. Mol. Cel/. Biol. 11: 4405, 1991.

24. Long, G. L., Chandrn, T., Woo, S. L. C., Davie, E. W., and Kura-chi, K. Complete sequence of the cDNA for human a,-anlitryp-sin and the gene for the S variant. Biochemistry 23: 4828, 1984.

25. Soriano, P., Friedrich, G., and Lawinger, P. Promoter interac-tions in retrovirus vectors introduced into fibroblasls and embry-onic stem cells. J. Virol. 65: 2314, 1991.

26. Markowitz, D., Goff, S., and Bank, A. A Bofe packaging for genetransfer: Separating viral genes on two different plasmids. J.Virol. 62: 1120, 1988.

27. Miller, A. D., and Rosman, G. J. Improved retroviral vectors forgene transfer and expression. Biotechniques 7: 980, 1989.

28. Fernandez-Lima, J. L., Moneo, l., Grubb, A., and Mendez, E. J.Immunol. Methods 82: 101, 1985~

29. Deleted in proof.

1. Starzl, T. E., Demetri!!, A. J., and van Thiel, D. I~iver tran!!plnn-tation. N. Engl. J. Med. 321: 1014, 1989.

2. Chowdhury, J. R., Gro!!sman, M., Gupta, S., Chowdhury, N. R.,Baker, J. R., and Wilson, J. M. Long-term improvement of hypcr-cholc!!tcrolcmia after ex vivo gene therapy in LDI~R-dcficiclltrabbit!!. Science 254: 1802, 1991.

3. Kay, M. A., Balcy, P., Rothenberg, S., Leland, F., Fleming, L.,Pondcr, K. P., Liu, T. T., Finegold, M., Darlington, G., Pokorny,W., and Woo, S. L. C. Expression of human a.-antitrypsin illdOK!! after autologous transplantation of retroviral transducedhepatocytes. Frac. Natl. Acad. Sci. USA 89: 89, 1992.

4. Kaneda, Y., Iwai, K., and Uhida, T. Introduction and expressionof the human insulin gene in adult rat liver. J. Biol. Chem. 264:12126, 1989.

5. Wu, G. Y., and Wu, C. H. Receptor mediated gene delivery andexpression in vivo. J. Biol. Chem. 263: 14621, 1988.

6. Hatzoglou, M., Lamers, W., Bosch, F., Wynshaw-Boris, A.,Clapp, D. W., and Hanson, R. W. Hepatic gene transfer in ani-mals using retroviruses containing the promoter from the genefor phosphoenolpyruvate carboxykinase. J. Biol. Chem. 265:17285, 1990.

7. Kay, M. A., Li, Q., Liu, T. J., Leland, F., Toman, C., Finegold,M., and Woo, S. L. C. Hepatic gene therapy: Persistent expres-sion ofhuman a1-antitrypsin in mice after direct gene delivery invivo. Hum. Gene Ther. 3: 641, 1992.

8. Kaleko, M., Garcia, J. V., and Miller, A. D. Persistent gene ex-pression after retroviral gene transfer into liver cells in vivo.Hum. Gene Ther. 2: 27, 1991.

9. Ferry, N., Duplessis, O., Houssin, D., Danos, O., and Heard, J. M.Retroviral mediated gene transfer into hepatocytes in vivo. Frac.Natl. Acad. Sci. USA 88: 8377, 1991.

10. Rozga, J., Moscioni, A. D., Neuzil, D., and Demetriou, A. A. Amodel for directed foreign gene delivery to rat liver cells in vivo.J. Surg. Res. 52: 209, 1992.

517HAFENRICHTER ET AL.: LIVER-SPECIFIC PROMOTER ELEMENTS

30. Deleted in proof.31. Yu, S. F., van Ruden, T., Kantoff, P. W., Garber, C., Seiberg, M.,

Ruther, U., Anderson, W. F., Wagner, E. F., and Gilboa, E. Self-inactivating retroviral vectors designed for transfer of wholegenes into mammalian cells. Proc. Nutl. Acud. Sci. USA 83:3194, 1986.

32. Yee, J. K., Moores, J. C., Jolly, D. J., Wolff, J. A., Respess, J. G.,and Friedmann, T. Gene express ion from transcriptionally dis-abled retroviral vectors. Proc. Nutl. Acud. Sci. USA 84: 5197,1987.

33. Soriano, P., Cone, P. D., Mulligan, R. C., and Jaenisch, R. Tissuespecific and ectopic expression of genes introduced into trans-genic mice by retroviruses. Science 234: 1409, 1986.

34. Weiss, R., Feich, N., Varmus, H., and Coffin, J. RNA TumorViruses Cold Spring Harbor, NY: Cold Spring Harbor Labora-tory Press, 1984.

35. Graves, B. J., Eisenman, R. N., and McKnight, S. L. Delineationof transcriptional control signals within the Moloney murinesarcoma virus long terminal repeat. Mol. Cell. Biol. 5: 1948, 1985.

36. Fabrikant, J. l. The kinetics of cellular proliferation in regener-ating liver. J. Cell. Biol. 36: 551, 1968.

37. Deleted in proof.38. Deleted in proof.39. Deleted in proof.

40. Deleted in proof.

41. Ponder, K. P., Gupta, S., Leland, F., Darlington, G., Finegold,M., DeMayo, L., Ledley, F., Chowdhury, J. R., and Woo, S. L. C.Mouse hepatocytes migrate to the liver parenchyma and func-tion indefinitely after intrasplenic transplantation. Proc. Natl.Acad. Sci. USA 88: 1217, 1991.

42. Lai, E., and Darnell, J. E. Transcriptional control in hepato-cytes: A window ondevelopment. Trends Biol. Sci. 16: 427,1991.

43. Johnson, P. F. Transcriptional activators in hepatocytes. CellGrowth Dilfer. 1: 47, 1990.

44. Pinkert, C. A., Ornitz, D. W., Brinster, R. L., and Palmiter, D. L.An albumin enhancer located 10 kb upstream functions alongwith its promoter to direct efficient, liver specific expression intransgenic mice. Genes Dev. 1: 268, 1987.

45. Jaffe, H. A., Daniel, C., Longnecker, G., Metzger, M., Setokuchi,Y., Rosenfeld, M. A., Gant, T. W., Thorgeirsson, S. S., Stratford-Perricaudet, L. D., Perricaudet, M., Pavirani, A., Lecocq, J.-P.,and Crystal, R. G. Adenovirus-mediated in vivo gene transferand expression in normal rat liver. Nature Genet. 1: 372, 1992.

46. Jainchill, J. L., Aaronson, S. A., and Todaro, G. J. Murine sar-coma and leukemia viruses: Assay using clonallines of contact-inhibited mouse cells. J. Virol. 4: 549, 1969.

47. Darlington, G. J. Liver celllines. Methods Enzymol. 151: 19,1987.