In Utero Delivery of Adeno-Associated Viral Vectors: Intraperitoneal Gene Transfer Produces Long-Term Expression

ARTICLE

doi:10.1006/mthe.2001.0267, available online at http://www.idealibrary.com on IDEAL

In Utero Delivery of Adeno-Associated Viral Vectors: Intraperitoneal Gene Transfer Produces Long-Term Expression Gerald S. Lipshutz,* Christopher A. Gruber,†,‡ Yu-an Cao,†,‡ Jonathan Hardy,†,‡ Christopher H. Contag,†,‡ and Karin M. L. Gaensler§,1 *Department of Surgery and §Department of Medicine, University of California, San Francisco, San Francisco, California 94143-0793 † Department of Pediatrics and ‡Department of Microbiology & Immunology, Stanford University, Stanford, California 94305-5208 Received for publication November 8, 2000; accepted in revised form January 10, 2001.

Recombinant adeno-associated viruses (rAAV) are promising gene transfer vectors that produce long-term expression without toxicity. To investigate future approaches for in utero gene delivery, the efficacy and safety of prenatal administration of rAAV were determined. Using luciferase as a reporter, expression was assessed by whole-body imaging and by analysis of luciferase activity in tissue extracts, at the time of birth and monthly thereafter. Transgene expression was detected in all injected animals. Highest levels of luciferase activity were detected at birth in the peritoneum and liver, while the heart, brain, and lung demonstrated low-level expression. In vivo luciferase imaging revealed persistent peritoneal expression for 18 months after in utero injection and provided a sensitive whole-body assay, useful in identifying tissues for subsequent analyses. There was no detectable hepatocellular injury. Antibodies that reacted with either luciferase or rAAV were not found. AAV sequences were not detected in germ-line tissues of injected animals or in tissues of their progeny. In utero AAV-mediated gene transfer in this animal model demonstrates that novel therapeutic vectors and strategies can be rapidly tested in vivo and that rAAV may be developed to ameliorate genetic diseases with perinatal morbidity and mortality. Key Words: in utero; gene therapy; adeno-associated virus; visible bioluminescence; intraperitoneal; luciferase; fetal.

INTRODUCTION Correction of genetic diseases during fetal development may be achieved by targeting genetic therapies to expanding stem cell populations, and to developing organ systems. Early gene transfer and long-term expression of therapeutic proteins during fetal development may limit or abrogate the pathologic consequences of genetic mutations. Lack of development of inhibitory immune responses may allow for successful postnatal readministration and expression of potentially immunogenic therapeutic genes (1). While these prenatal therapeutic strategies are attractive, in utero gene transfer approaches

1 To whom correspondence and reprint requests should be addressed at Department of Medicine, 3rd and Parnassus Avenues, Box 0793, University of California, San Francisco, CA 94143-0793. Fax: (415) 5664969. E-mail: [email protected].

284

must first be carefully evaluated to ensure that there is no significant morbidity to the fetus or mother, and no alteration of germ-line tissues (2, 3). Using different animal models, several groups have delivered adenoviruses (4 – 8), retroviruses (8), or plasmid vectors (9) to the developing fetus. Amniotic fluid instillation has been performed to transduce cells of the respiratory and gastrointestinal systems (5– 8). Delivery of an adenoviral vector by umbilical vein puncture in fetal sheep resulted in detectable transgene expression in many fetal tissues with predominance in the liver (10). Intraplacental introduction of cellular grafts have produced foreign gene products throughout gestation (11). Direct intracardiac (12), intraperitoneal (ip) (13), or intrahepatic (14) injection of adenoviral vectors containing either reporter (13, 14) or therapeutic genes (13) has produced gene expression throughout the neonatal period in murine models. These successful studies have shown that in MOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy 1525-0016/01 $35.00

ARTICLE utero gene transfer is possible, and have demonstrated the need to develop strategies that provide long-term gene expression without pathologic consequences. As an in utero gene delivery vector, recombinant adenoassociated virus 2 (rAAV) would have several significant advantages. A key advantage of using this vector is that AAV-mediated gene delivery may produce, long-term gene expression following prenatal administration as it has after administration in adult mammalian tissues (15– 18). Additional advantages of rAAV as gene transfer vectors include (i) a broad host cell range (19), (ii) an ability to infect dividing and growth-arrested cells (20), (iii) the availability of procedures for high titer, helper virus-free preparation (21, 22), and (iv) a lack of toxicity and reduced immunogenicity (20). These features of rAAV vectors suggest that they may be well-suited for prenatal gene transfer, and should be thoroughly evaluated in appropriate animal models. We evaluated the efficacy of AAV-mediated gene transfer and the potential for adverse effects following a single transuterine, intraperitoneal injection of AAV-EF1␣-luciferase into murine fetuses at day 15 of gestation. We assessed the survival and development of treated pups, and evaluated transgene expression patterns using sensitive in vivo (23, 24) and ex vivo assays for luciferase activity. We also investigated whether treated pups developed humoral immune responses to rAAV, or to luciferase, as neutralizing antibody responses have been observed in adult mice that received rAAV-based vectors via ip injection (25, 26). Analysis of the duration and tissue distribution of transgene expression revealed long-term expression without evidence of acute or chronic toxicity in animals treated in utero. AAV transduced cells in both intraabdominal and more distant tissues and directed long-term expression in the peritoneum.

RESULTS Survival, Growth, and Development of Animals Injected with AAV in Utero Fetuses of 5 pregnant CD-1 females underwent transuterine injection at E15 (embryonic day 15) with 3 ⫻ 1011 genomes of rAAV containing the EF1␣ promoter and a luciferase reporter gene (AAV-EF1␣-luciferase). A total of 53 fetuses received in utero injections. One female was euthanized on E18. All of the E18 fetuses (n ⫽ 8) had similar weights (1.11 ⫾ 0.08 grams vs 1.15 ⫾ 0.05 grams, P ⫽ 0.39) and similar crown-to-rump lengths (20.1 ⫾ 0.9 mm vs 20.4 ⫾ 0.9 mm, P ⫽ 0.69) when compared to age-matched, uninjected controls. Forty-one of 45 of the remaining pups were live born (91.1%) and showed normal growth and development. This compares favorably with the live-born rates of mock-injected pups (81–100%) (14). MOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy

FIG. 1. Level of luciferase activity in newborn tissues at day 1 of life after transuterine ip injection of AAV-EF1␣-luciferase (n ⫽ 4). X axis indicates relative light units/␮g tissue protein; Y axis represents individual tissues (1 ⫽ brain, 2 ⫽ heart, 3 ⫽ lung, 4 ⫽ intestine, 5 ⫽ liver, 6 ⫽ kidney, 7 ⫽ spleen, 8 ⫽ thymus, 9 ⫽ uninjected control). (Error bars represent SEM.)

Tissue Distribution of Adeno-Associated Viral Expression All of the mice that received in utero injections showed high levels of luciferase expression. Very high levels of luciferase activity were detected in the tissues extracts from the livers of 1-day-old pups (n ⫽ 4) (72,548 ⫾ 7509 RLU/␮g protein). High levels of expression were also present in the heart (4045 ⫾ 3536 RLU/␮g protein). Lower levels of luciferase activity were detected in all remaining organs examined, including the brain (169 ⫾ 151 RLU/␮g protein), intestine (907 ⫾ 741 RLU/␮g protein), kidney (381 ⫾ 252 RLU/␮g protein), lung (441 ⫾ 423 RLU/␮g protein), spleen (444 ⫾ 219 RLU/␮g protein), and thymus (18 ⫾ 5 RLU/␮g protein) (Fig. 1). Uninjected tissues and cell lines transduced in vitro with rAAV carrying other reporter genes demonstrated background levels of 0.05 ⫾ 0.004 RLU/␮g protein. To determine whether hematopoietic progenitors residing in the fetal liver were efficiently transduced by midgestation ip administration of rAAV, PCR analyses of peripheral blood from 3-week-old and 8-month-old animals was performed. There were no positive PCRs from 1.2 ⫻ 106 genome equivalents (8 ␮g of genomic DNA) obtained from whole blood of each of three 8-month-old animals. No luciferase sequences were detected from 0.6 ⫻ 106 genome equivalents (4 ␮g of genomic DNA) obtained from whole blood of each of three 3-week-old animals. Hence, recombinant AAV genomes could not be detected in approximately 1 ⫻ 106 nucleated blood cells obtained from either 3-week-old or 8-month-old mice that were transduced in utero with rAAV-EF1␣-luciferase.

Duration of Luciferase Levels after in Utero Injection Luciferase expression was detected, via the ex vivo enzyme assay, in the liver at all time points examined. Hepatic luciferase activity was highest at birth (72,548 ⫾

285

ARTICLE

FIG. 2. Duration of luciferase expression in selected tissues. Tissues were harvested from four animals at birth (n ⫽ 4) and monthly thereafter (n ⫽ 3). Error bars represent SEM.

7,509 RLU/␮g protein) and then declined during the rapid early phase of growth during the first 4 weeks of life (Fig. 2). Hepatic luciferase expression levels stabilized after the first month of life, and persisted at low levels, throughout the study period (97 ⫾ 16 RLU/␮g protein at 6 months) (Fig. 2). Luciferase expression in the heart and brain also declined over the first 4 weeks of life, but remained stable thereafter (heart, 4045 RLU/␮g protein at birth and 44 RLU/␮g protein at 6 months; brain, 187 RLU/␮g protein at birth and 7 RLU/␮g protein at 6 months) (Fig. 2).

In Vivo Analysis of Bioluminescence in Tissue and Organ Sites The tissue distribution of luciferase expression was analyzed in mice (n ⫽ 4) previously injected in utero using in vivo real-time imaging of bioluminescence on E18, day 4 of life, and at 1, 2, 3, 4, 6, 10, 12, 13, 14, 17, and 18 months of age (Figs. 3 and 4). Luciferase expression was detected 3 days after in utero injection (Fig. 3A) and remained stable in animals up to 18 months of age (end of study) (Fig. 3F). In these animals, expression of luciferase in abdominal and distal organs was obscured by the very high-level expression in the peritoneal tissues overlying these organs (Figs. 4A– 4C). In vivo monitoring of animals after 1 month of age demonstrated that in most cases, luciferase expression was limited to the peritoneum. Occasionally, persistent luciferase activity was observed in the liver and brain, suggesting that the rAAV vector was delivered intravenously. An initial decline in expression, as determined by quantitation of total photons emitted, was observed during the first month of life, consistent with the decline in luciferase activity in tissue extracts demonstrated by luminometry (Fig. 5). Although the peritoneal surface area increased approximately 22-fold from birth to adulthood (133 ⫾ 4 mm2 at birth to 2881 ⫾ 295 mm2 at adult size), peritoneal expression of luciferase at 18 months remained high. Thus was demonstrated both by quantitation of total photons emitted using in vivo bioluminescence (Fig. 5), and by luminometry (at 10 months, right-sided peritoneal expression was 9271

286

RLU/␮g protein and left-sided peritoneal expression was 7,956 RLU/␮g protein) (Fig. 4A). To determine the sensitivity of the in vivo luciferase assay, three animals with different patterns of reporter gene expression were selected from the animals that received AAV-EF1␣-luciferase in utero. At 8 months of age, reporter gene expression levels in these animals were analyzed both in vivo, and by ex vivo assays that assess reporter protein levels (Western analyses) and luciferase DNA copy number (Fig. 6). Liver tissue was selected for these analyses as the peritoneum was difficult to reproducibly dissect free of adjoining muscle tissue. In vivo luciferase assays could detect 38 pg of luciferase in 1 g of liver tissue, corresponding to approximately 1 ⫻ 109 molecules of luciferase in the liver. This level of luciferase was expressed from approximately 1 rAAV genome in 1 ⫻ 106 mouse genome equivalents as determined by PCR on diluted mouse liver DNA. Thus, in vivo luciferase imaging detected 38 pg of luciferase per gram of liver tissue indicating that whole-body imaging was as sensitive as ex vivo luciferase assays and more sensitive than Western blot analyses.

Absence of Hepatic Toxicity, Humoral Immune Responses, or Detectable Germ-Line Integration Following in Utero Delivery of rAAV Mice that had undergone in utero ip injection with recombinant AAV produced litters that were comparable in size to those of age-matched controls (mean 12.4 pups/ litter and 11.1 pups/litter, respectively). Serum alanine aminotransferase (ALT) levels, a sensitive measure of hepatocellular injury, were similar in samples from animals injected with rAAV in utero and from age-matched uninjected controls at 1, 2, 3, 4, 5, and 6 months of age. At 3 months of age, the ALT level of rAAV-injected mice was 48 ⫾ 3.6 IU/ml and the ALT level of uninjected mice was 47.3 ⫾ 4.2 IU/ml (data from other time points not shown). None of the mice injected ip with AAV-EF1␣luciferase in utero developed obvious health problems or died prematurely. When adult animals were dissected for harvest of tissues at different time points, there was no evidence of peritonitis, or of adhesions suggestive of an inflammatory reaction. No mononuclear infiltrate consistent with a cellular immune response was detected in the livers from animals 1– 6 months of age by examination of H & E-stained sections (data not shown). Serum samples from uninjected and in utero injected animals were analyzed by ELISA for anti-luciferase antibodies at 1, 3, and 5 months after in utero injection and for anti-AAV-2 antibodies at 5 months. No antibodies either to luciferase (data at 5 months shown in Fig. 7A; 1 and 3 months, data not shown) or to AAV-2 (data at 5 months shown in Fig. 7B) were detected in animals injected in utero. PCR amplification performed to determine whether germ-line tissues were transduced, did not detect AAVEF1␣-luciferase vector sequences in germ-line tissues (testes n ⫽ 6, ovaries n ⫽ 6) of adult animals injected in utero, or of their pups (n ⫽ 217). Positive controls included MOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy

ARTICLE

FIG. 3. In vivo monitoring of bioluminescence. Before imaging, individual animals (identified by numbers) were placed at the identical stage position. A pseudocolor reference range is located to the right of each figure with maximum expression displayed above. Animals were imaged (A) in utero, expression by visible bioluminescence is detected in three fetuses, (B) day 4 of life, (C) 1 month, (D) 6 months, (E) 12 months, and (F) 18 months of age. Reference range is the same for images at 1 through 18 months. FIG. 4. In vivo monitoring of bioluminescence in animals at 10 months of age. (A) The mouse is imaged in the supine position. (B) An upper transverse incision has been made and the skin reflected caudally (in direction of arrow). (C) The muscle and peritoneum have been incised and are reflected cranially (in direction of arrow) exposing the abdominal viscera below. R, right; L, left side of animal; A, adipose tissue; I, intestine; LA, linea alba; S, spleen; U, umbilicus.

tissues from adult mice after AAV delivery, and genomic DNA samples from uninjected control mice to which known amounts of the AAV-EF1␣-luciferase plasmid were added. Consistent with this finding, hepatic luciferase expression in these progeny (0.03 ⫾ 0.003 RLU/␮g protein) was comparable to that of control background levels [0.05 ⫾ 0.03 RLU/␮g protein (n ⫽ 15)].

DISCUSSION These studies demonstrate that intraperitoneal delivery of AAV in the fetal mouse produces high-level gene expresMOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy

sion in the neonatal period. All of the mice injected in utero expressed high levels of luciferase during the first week of life. The duration of expression extended well into adult life. Expression of the transferred gene was detected adjacent to the site of injection in the peritoneum, and at lower levels in intra-abdominal organs. Tissue sites distant from the peritoneal cavity were also transduced. In 1-month-old animals, highest levels of luciferase expression were observed in the peritoneum both by in vivo whole body bioluminescence imaging, and by image-guided luminometric analyses of tissue extracts. Postnatal growth and development were normal, and vec-

287

ARTICLE

FIG. 5. Quantitation of total bioluminescence from animals previously injected in utero at 4 days, 2 weeks, 1 month, 2 months, 3 months, 10 months, 12 months, 13 months, 14 months, 17 months, and 18 months—X axis (n ⫽ 3– 4 at each time point). Y axis indicates total photons emitted.

tor sequences were not detected in germ-line tissues. There was no detectable humoral immune response to the luciferase reporter gene product, or to rAAV-2 in animals receiving the vector in utero. This is in contrast to the humoral immune responses observed with rAAV delivery in adult animals (1, 25, 26). We have recently demonstrated that delivery of highly immunogenic adenoviral vectors to day 15 murine fetuses did not elicit humoral or lymphoproliferative responses to either the transferred gene product or to adenoviral antigens (27). Readministration of recombinant adenovirus to mice previously injected with this virus in utero produced levels of gene expression comparable to those of age-matched uninjected controls. However, both naive mice and adult animals injected once in utero, developed brisk immune responses to adenoviral antigens and to luciferase after intravenous adenoviral delivery as adults. The absence of immune responses in animals exposed to viral gene delivery vectors in utero probably reflects the immaturity of the immune system throughout gestation and the neonatal period previously described in fetal/ neonatal transplantation studies (28, 29). In utero gene transfer studies performed in small and large animals have generally demonstrated short-term gene expression, with the exception of retroviral-mediated gene transfer studies that produced long-term, lowlevel expression up to 5 years (4 –14, 30). However, oncoretroviral vectors only transduce dividing cells and are

288

largely inactivated in serum. rAAV has the advantage, therefore, of both long term expression and transduction of non-dividing cells (15, 16). In our previous studies, in utero delivery of adenoviral vectors containing the murine factor VIII gene produced therapeutic levels of factor VIII that persisted throughout the neonatal period in mice with hemophilia A (13). However, adenoviral-mediated gene expression was not detected after mice reached 3 weeks of age. The efficiency of peritoneal transduction and the persistence of gene expression suggests that ip administration of rAAV in utero may be an efficient strategy for the long-term production of secreted therapeutic proteins such as coagulation factors or ␣1-antitrypsin. We were unable to detect any pathologic effects despite persistent expression of the transgene after in utero delivery of high titer rAAV (3 ⫻ 1011 genomes). Fetal survival rates (91%) were comparable to those described in our previous studies using adenoviral vectors (4, 13, 14). An initial decline in expression, as determined by quantitation of total photons emitted, was observed during the first month of life. This is consistent with the decline in luciferase activity in tissue extracts demonstrated by luminometry. There are a variety of mechanisms by which AAV-mediated gene expression may decline over time. The decremental expression is at least in part due to animal growth from 18 days of gestation to adult size and associated dilution of episomal AAV vector sequences. In addition, host cell mediated silencing of integrated AAV genomic sequences may be an important mechanism underlying decremental gene expression. Finally, the extent to which the AAV vector remains in an episomal form, or integrates into the host genome is an important variable. Loss of episomal AAV genomes during rapid cycles of cell division could also be an underlying factor in the progressive loss of luciferase activity in transduced tissues. Although the liver was transduced in utero, it is unlikely that significant numbers of hematopoietic stem cells were transduced. PCR analyses of peripheral blood samples from 3-week-old and 8-month-old mice were uniformly negative as they did not demonstrate the presence of luciferase sequences. This suggests a lack of efficient stable transduction of hematopoietic progenitors accounts for the absence of luciferase in blood rather than transcriptional silencing of rAAV expression mediated by the host cells. The identification of individual cells that are transduced and expressing luciferase (luciferase-positive and PCR-positive), versus those cells that are transduced by AAV where gene expression has been silenced (PCR-positive only), is difficult to accurately assess in solid organs or tissues such as the peritoneum, particularly when luciferase is the reporter gene. In adult animals, AAV-mediated gene expression reaches peak levels several weeks after injection (31, 32). The slow rise in AAV-mediated gene expression appears to be due to the conversion of single-stranded monomer genomes into double-stranded, high-molecular-weight, head-to-tail concatamers occurring over a period of five weeks after intraportal and intramuscular rAAV administration (31, 32). In contrast to the delayed onset of exMOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy

ARTICLE

FIG. 6. Sensitivity of luciferase detection in vivo. (A) In vivo bioluminescence. Three animals that received AAV in utero were selected based on their different in vivo luciferase expression patterns, and compared to an uninjected control animal. The in vivo patterns of luciferase expression indicated that one of the injected animals had unilateral peritoneal expression and no detectable signal over the liver (AAV1), one animal presented with the more typical bilateral peritoneal expression and weak signal from the liver (AAV4), and one animal presented with the unusual pattern of primarily liver expression. (B) Ex vivo liver bioluminescence. Luciferase signal was detected in excised livers of animals AAV4 and AAV7 both in intact organs and in tissue lysates. In comparison to a standard curve of luciferase activity, 38 and 445 pg of luciferase protein per gm of liver tissue was detected in animals AAV4 and AAV7, and no signal was detected in the control or AAV1. By Western blot analyses a detectable signal was apparent only in AAV7; the sensitivity of the Western blot detection was 10 pg of luciferase in 25 ␮g of liver protein (data not shown). (C) PCR quantitation of AAV-EF1␣-luciferase copy number per genome equivalent. Three of 20 separate PCRs, each containing 1 ␮g of liver DNA (1.5 ⫻ 105 genome equivalents) from animal AAV4, were PCR positive, indicating a copy number of 1 in 1 ⫻ 106 genome equivalents. The copy number in animal AAV7 was estimated to be 1 in 6 ⫻ 103 genome equivalents, and no signals were detected in either the control or animal AAV1. Thus, in vivo luciferase imaging detected 38 pg of luciferase per gm of liver tissue indicating that whole-body imaging was as sensitive as ex vivo luciferase assays and more sensitive than Western blot analyses.

pression observed in adult animals, rAAV expression was detected within a few days after fetal injection. A significant proportion of early luciferase expression may be directed from episomal rAAV sequences that are diluted or lost as the animal grows to adult size. In rapidly growing fetal tissues, cells undergoing DNA replication and cell division are more abundant than in most adult tissues. Thus, the developing fetus may provide a milieu where the AAV genome is more rapidly converted to the doublestranded, concatameric form and expressed. Studies are currently underway to define the ratio of integrated to episomal rAAV genomes in fetal and adult tissues. MOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy

The use of real-time, in vivo monitoring of bioluminescence makes longitudinal studies of levels and tissue distribution of luciferase expression in mice feasible, and offers improved statistical power through repeated measurements (23, 24, 33). In biochemical assays, when luciferin is added and luciferase activity analyzed, the luminescence rises to a maximum intensity in approximately 2 s and declines with a t1/2 of about 10 s (34). Luciferase has a half-life of approximately 3 h in cell culture and in vivo. This facilitates temporal assessment of gene expression, in vivo, since the reporter signal will accurately reflect both induction and reduction of transcription (35).

289

ARTICLE (39). Third, fetal immune immaturity results in no detectable immune response to the vector or the transgene, and may induce immune tolerance (39). Although not evaluated in this study, the long term persistence of luciferase expression following AAV gene transfer may be more conducive to induction of immune tolerance than the short-term, high-level gene expression produced by adenoviral vectors. Finally, the development of ultrasoundguided techniques, currently used for peripheral umbilical blood sampling and prenatal transfusions, greatly limits the risk to the fetus and mother. The efficacy and safety of rAAV delivered prenatally demonstrates the potential of these vectors for in utero gene transfer and for the development of future therapeutic strategies.

EXPERIMENTAL PROTOCOL

FIG. 7. Analysis of antibody responses to luciferase (A) and rAAV (B). Serum from 5-month-old mice injected in utero (n ⫽ 3) (black bars); uninjected controls (n ⫽ 3) (white bars). Error bars represent SD.

Bioluminescent markers have low background levels and are superior to fluorescent reporters where tissue autofluorescence may limit sensitivity. The sensitivity of detection in adult mice in this study was 38 pg of luciferase enzyme in 1 g of liver tissue, which was near the detection limits of ex vivo luciferase assays. Transgene expression profiling using whole body imaging was sensitive and provided more information more rapidly than did ex vivo assays of luciferase activity. This enabled a spatiotemporal analysis of transferred gene expression, directed our analyses to the peritoneal wall and liver, and thus was useful for directing the ex vivo assays to specific tissue sites. In vivo bioluminescence was detected at high levels despite animal growth, increased body density, and attenuation of signal due to absorbance and scattering of light by hair and tissues. The efficient transduction and longterm expression in the peritoneal membrane may be due to the higher relative multiplicity of infection of the virus to which this tissue is exposed with intraperitoneal administration of high titers of rAAV. The presence of luciferase activity in many tissues at birth suggests that the virus enters the circulation after an intraperitoneal injection in utero and is distributed systemically. Viral particles may travel through stomata located in the sub-diaphragmatic peritoneum (36), which are large enough to permit passage of AAV. Gene transfer to the developing fetus offers several potential advantages over postnatal gene transfer. First, in utero gene therapy may allow for the correction or prevention of genetic diseases before the onset of pathologic processes (37). Second, highly proliferative fetal stem cells may be more efficiently transduced than are more quiescent adult stem cell populations (38). Organs inaccessible later in life may also be more readily transduced in utero

290

Preparation of recombinant adeno-associated viral vectors. AAV-EF1␣-luciferase is a serotype 2 adeno-associated viral vector containing the firefly luciferase (luc) reporter gene (40). To construct this vector, pSSV9, a rAAV construction plasmid (provided by R. J. Samulski, University of North Carolina), was digested at XbaI releasing the complete viral genome (replication and capsid genes) and leaving the inverted terminal repeats. A fragment containing the EF1␣ promoter (41), a multicloning site, and the bovine growth hormone polyadenylation signal (42) was excised from pEF6/V5-His (Invitrogen, Carlsbad, CA) and cloned into pSSV9. An NcoI– XbaI fragment containing the luciferase cDNA was excised from pGL3enhancer plasmid (Promega, Madison, WI) and subcloned into the multicloning region, creating pSSV9-EF1␣-luciferase. This plasmid was used with pXX2 and pXX6 plasmids (21) (both provided by R. J. Samulski, University of North Carolina) to prepare rAAV stocks, as described (22). Viral titer was estimated by dot-blot hybridization. Animal procedures. All procedures were approved by the University of California, San Francisco (UCSF) Committee on Animal Research. CD-1 female and male mice were purchased from Charles River Breeding Laboratories (Wilmington, MA). Timed matings were established and vaginal plug dates recorded as day 1 of gestation. On day 15 of gestation, transuterine, intraperitoneal injection of 3 ⫻ 1011 genomes of rAAV in 10 ␮l of normal saline was performed as previously described (4). Control animals were not injected. Ex vivo assay of luciferase activity. Animals were euthanized, individual organs harvested, and tissue extracts prepared as described (4). Luciferase activity was determined in triplicate by luminometry (Monolight 3010) (Analytical Luminescence Laboratory, Sparks, MD) as previously described (4). Relative light units (RLU) per sample was calculated after adjusting for background activity in samples of lysis buffer alone. Samples were analyzed with a Lowry-based protein assay (43) and luciferase activity was plotted for each tissue as RLU/␮g tissue protein as described (4). The amount of luciferase protein (pg) per gram of tissue (wet weight) was estimated in selected liver samples based on the RLU per nanogram of purified luciferase enzyme (Promega). Imaging of luciferase activity in vivo. Mice were anesthetized (4) and an aqueous solution of luciferin substrate (150 ␮g/g body wt) (Biosynth A.G., Switzerland) was injected into the peritoneal cavity 5 min before beginning imaging. Intraperitoneal injection of luciferin results in rapid uptake into the vascular system and yields similar data, but slightly delayed light emission, to that obtained with intravenous delivery of the substrate (data not shown). Mice were placed in a light-tight chamber and a gray-scale reference image was obtained under low-level illumination (33). In complete darkness, photons transmitted through the tissues were collected using an intensified charge-coupled device (ICCD) camera (Model C240032: Hamamatsu Photonics, Hamamatsu City, Japan), with a 5-min integration time. A pseudocolor image representing light intensity (purple lowest intensity and red greatest intensity) was generated on an Argus 20 image processor (Hamamatsu Photonics, Hamamatsu City, Japan). Images were collected using LivingImage software (Xenogen Corp., Alameda, CA), MOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy

ARTICLE a software overlay for the image analysis program Igor (Wavemetrics, Seattle, WA). The relative light intensity from the whole animal, or from areas of interest, was quantified using the image analysis capabilities of LivingImage and Igor. Determination of serum alanine aminotransferase (ALT). Blood was drawn from the retro-orbital plexus in adult mice and analyzed as described (4). The results were pooled and mean and SEM calculated. Histological examination. Right lobes of livers from experimental and control animals 1– 6 months old were processed as described (44) and photographed at 400⫻ magnification. ELISA for anti-luciferase antibodies. Ninety-six-well plates were coated with 0.625 ␮g/ml of recombinant luciferase (Promega) in 100 ␮l of carbonate buffer. Plates were incubated at 37°C for 2 h, and then washed with PBS. Serum samples were diluted 1/50 in blocking buffer (PBS with 1% BSA and 0.05% Tween 20). A 1:2 serial dilution was prepared from 1:40 to 1:5120 and added to plates and incubated at 37°C for 1 h. After washing with PBS, a goat anti-mouse peroxidase-conjugated secondary antibody (Gibco/BRL, Rockville, MD) was prepared at a 1:500 dilution in blocking buffer, and 50 ␮l was added to each well. Plates were incubated for 1 h at 37°C, washed with PBS, and 50 ␮l of OPD substrate (Sigma, St. Louis, MO) was added to each well. After incubating for 10 min at 20°C, the reaction was stopped by adding 50 ␮l of 2.5 M H2SO4 and the plates read at 492 nm. All samples were assayed in duplicate, and an average and standard deviation (SD) calculated for each dilution. The OD value of the background signal (blocking buffer without serum) was subtracted from sample values. Samples were considered positive when the OD was at least threefold higher than the negative control. Positive control sera were obtained from serum samples of adult mice that had been injected with adeno-luciferase (4) and had previously titered anti-luciferase antibodies using this assay. Three animals injected prenatally were tested at 1, 3, and 5 months. ELISA for anti-AAV-2 antibodies. Ninety-six-well plates were coated overnight at 4°C with 1 ⫻ 108 genomes of purified rAAV/50-␮l well in PBS. An ELISA was then performed as outlined above. Positive control sera were obtained from AAV-EF1␣-luciferase-injected adult mice that had previously been tested for AAV antibodies using this assay. rAAV-injected (n ⫽ 3) and control animals (n ⫽ 3) were tested at 5 months of age. Analysis of germline tissues and determination of rAAV copy number in liver and blood. Gonads of male (n ⫽ 6) and female (n ⫽ 6) mice previously injected in utero were analyzed following PCR amplification. In addition, progeny of animals injected in utero were euthanized in the first week of life and their livers removed and divided in half. One-half of the liver was used for genomic DNA isolation and PCR amplification. The second hepatic fragment was used for the luciferase assay. Luciferase primers amplified a 369-bp luc gene fragment (45). Samples were resolved in 0.8% Nusieve 3:1 agarose (FMC BioProducts, Rockland, ME) in TBE. Mouse ␤-actin primers were used as internal controls (13). Vector standards for determining sensitivity of PCR were generated by adding 50, 5, 0.5, 0.05, 0.005, 0.001, and 0.0005 ng of pSSV9-EF1␣-luciferase plasmid DNA to 500 ng of mouse genomic DNA. The limits of detection in this assay was 5 pg of transgene DNA in 500 ng of genomic DNA. Positive controls included genomic DNA spiked with pSSV9-EF1␣-luciferase plasmid DNA and genomic DNA from adult tissues that had previously been injected with AAV-EF1␣-luciferase. Genomic DNA was prepared (DNeasy Tissue Kit, Qiagen, Valencia, CA) from citrated whole blood of mice at 3 weeks and 8 months of age and liver at 8 months. To estimate the number of luciferase DNA copies in mouse liver DNA, two rounds of PCR were employed using primers luc-1 (5⬘ATGAATTCAAGCTTATGGAAGACGCCAAAAA CATAAAGAAAGG CCCGGCGCCATTCTATC-3⬘; corresponding to positions 1– 46 of the luciferase coding sequence) and luc-2 (5⬘-CTGATCCTTAGGTACCCCTCCACCCAGCCCGCCTGAACCTCCCACG GCGATCTTTCCGCCCTTCTTGGCCTTTATGAGGATC-3⬘; corresponding to positions 1611–1650) as first round primers and Luc-3 (5⬘-ATGAACTCCTCTGGATCTACTGGTCTGC3⬘; corresponding to positions 586 – 613), and Luc-4 (5⬘-TAACACGGCGATCTTTCCGCCCTTCTTG-3⬘; corresponding to positions 1626 –1651) in the second round. First round conditions were 94°C for 5 min; 40 cycles of 94°C for 30 s, 68°C for 30 s, 72°C for 2 min and 30 s. Second round conditions were 94°C for 5 min; 40 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 2 min. Single copies of luciferase DNA in 1 ␮g of mouse MOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy

genomic DNA could be detected with this primer set. rAAV-luciferase DNA copy number were estimated by limiting dilution (10-fold and then 2-fold dilution) of tissue-derived DNA in control DNA. The last dilution that had at least two positive reactions of three were considered the dilution with a single copy. Using the relationship of 1 ␮g of liver DNA represents approximately 1.5 ⫻ 105 mouse genome equivalents, the DNA copy numbers per cell number were estimated for liver tissue of selected animals. 0.5 ␮g of DNA from whole blood of 3-week-old mice (n ⫽ 3) and 1.0 ␮g of DNA from whole blood and liver of 8-month-old mice underwent nested PCR (8 –10 separate reactions/sample). Positive controls included samples with 1 copy and 5 copies of luciferase cDNA. Western analysis. Polyclonal rabbit anti-luciferase antibody (Cortex Biochem, San Leandro, CA) was used to estimate levels of firefly luciferase in mouse liver, following size separation of 25 ␮g total protein by polyacrylamide gel electrophoresis (7.5% polyacrylamide) and transfer to solid support (Immunolon, Millipore). The second antibody was HRP-conjugated goat anti-rabbit IgG, which was detected with a chemiluminescent reagent (ECL, Amersham, Piscataway, NJ). Samples of liver lysate were normalized for protein content (Bio-Rad DC assay, Bio-Rad, Hercules, CA). A standard curve of purified luciferase enzyme (Promega) was used to determine amounts of luciferase enzyme per ␮g of liver protein. To ensure the detectability of luciferase in liver lysate, the luciferase was diluted in sample buffer, with 25 ␮g of negative control liver lysate protein added to each sample. Statistical calculations. Mean, standard deviation, standard error of the mean, and Student’s t test were calculated using standard formulae (46).

ACKNOWLEDGMENTS This work was supported in part by National Institutes of Health Grants RO1 HL62389-01 (K.M.L.G.) and 1RO1 HL58013 and 5RO1 HD37543-02 (C.H.C.), a fellowship from the Bank of America/Giannini Foundation (G.S.L.), a scholarship from the American College of Surgeons (G.S.L.), an unrestricted gift from the Dr. Harris M. Fishbon Fund of the Mt. Zion Health Fund (K.M.L.G., G.S.L.), the ONR, Contract N-00014-94-1-1024, and unrestricted gifts from the Mary L. Johnson and Hess Research Funds (C.H.C.). We thank Jennifer Wilson, Chin Lin, and Stephanie Andriole for skillful technical assistance and Dr. Jude Samulski for supplying plasmids pXX6, pXX2, and pSSV9. The authors gratefully acknowledge Dr. Y. W. Kan, Dr. Nancy L. Ascher, and Dr. David Lewis for their continued interest in this work. C.H.C. has a financial interest in and is a paid consultant for Xenogen Corp.

REFERENCES 1 Halbert, C. L., Rutledge, E. A., Allen, J. M., Russell, D. W., and Miller, A. D. (2000). Repeat transduction in the mouse lung by using adeno-associated virus vectors with different serotypes. J. Virol. 74: 1524 –1532. 2 Zanjani, E. D., and Anderson, W. F. (1999). Prospects for in utero human gene therapy. Science 285: 2084 –2088. 3 King, D., Shakespeare, T., Nicholson, R., Clarke, A., and McLean, S. (1999). Risks inherent in fetal gene therapy. Nature 397: 383. 4 Lipshutz, G. S., Flebbe-Rehwaldt, L., and Gaensler, K. M. L. (1999). Adenovirus-mediated gene transfer in the midgestation fetal mouse. J. Surg. Res. 84: 150 –156. 5 McCray, P. B., Jr., et al. (1995). Adenoviral-mediated gene transfer to fetal pulmonary epithelia in vitro and in vivo. J. Clin. Invest. 95: 2620 –2623. 6 Vincent, M. C., et al. (1995). Adenovirus-mediated gene transfer to the respiratory tract of fetal sheep in utero. Hum. Gene Ther. 6: 1019 –1028. 7 Holzinger, A., Trapnell, B. C., Weaver, T. E., Whitsett, J. A., and Iwamoto, H. S. (1995). Intraamniotic administration of an adenoviral vector for gene transfer to fetal sheep and mouse tissues. Pediatr. Res. 38: 844 – 848. 8 Douar, A. M., Adebakin, S., Themis, M., Pavirani, A., Cook, T., and Coutelle, C. (1997). Fetal gene delivery in mice by intra-amniotic administration of retroviral producer cells and adenovirus. Gene Ther. 4: 883– 890. 9 Gaensler, K. M. L., et al. (1999). Fetal gene transfer by transuterine injection of cationic liposome–DNA complexes. Nat. Biotechnol. 17: 1188 –1192. 10 Themis, M., et al. (1999). Successful expression of beta-galactosidase and factor IX transgenes in fetal and neonatal sheep after ultrasound-guided percutaneous adenovirus vector administration into the umbilical vein. Gene Ther. 6: 1239 –1248. 11 Senut, M. C., Suhr, S. T., and Gage, F. H. (1998). Gene transfer to the rodent placenta in situ. A new strategy for delivering gene products to the fetus. J. Clin. Invest. 101: 1565–1571. 12 Woo, Y. J., et al. (1997). In utero cardiac gene transfer via intraplacental delivery of recombinant adenovirus. Circulation 96: 3561–3569. 13 Lipshutz, G. S., Sarkar, R., Flebbe-Rehwaldt, L., Kazazian, H., and Gaensler, K. M. L. (1999). Short-term correction of factor VIII deficiency in a murine model of hemophilia A

291

ARTICLE after delivery of adenovirus murine factor VIII in utero. Proc. Natl. Acad. Sci. USA 96: 13324 –13329. 14 Lipshutz, G. S., Flebbe-Rehwaldt, L., and Gaensler, K. M. L. (1999). Adenovirusmediated gene transfer to the peritoneum and hepatic parenchyma of fetal mice in utero. Surgery 126: 171–177. 15 Xiao, X., Li, J., and Samulski, R. J. (1996). Efficient long-term gene transfer into muscle tissue of immunocompetent mice by adeno-associated virus vector. J. Virol. 70: 8098 – 8108. 16 Monahan, P. E., et al. (1998). Direct intramuscular injection with recombinant AAV vectors results in sustained expression in a dog model of hemophilia. Gene Ther. 5: 40 – 49. 17 Laughlin, C. A., Cardellichio, C. B., and Coon, H. C. (1986). Latent infection of KB cells with adeno-associated virus type 2. J. Virol. 60: 515–524. 18 McLaughlin, S. K., Collis, P., Hermonat, P. L., and Muzycka, N. (1988). Adenoassociated virus general transduction vectors: Analysis of proviral structures. J. Virol. 62: 1963–1973. 19 Kotin, R. M. (1994). Prospects for the use of adeno-associated virus as a vector for human gene therapy. Hum. Gene Ther. 5: 793– 801. 20 Grimm, D., and Kleinschmidt, J. A. (1999). Progress in adeno-associated virus type 2 vector production: Promises and prospects for clinical use. Hum. Gene Ther. 10: 2445–2450. 21 Xiao, X., Li, J., and Samulski, R. J. (1998). Production of high-titer recombinant adeno-associated virus vectors in the absence of helper adenovirus. J. Virol. 72: 2224 –2232. 22 Matsushita, T., et al. (1998). Adeno-associated virus vectors can be efficiently produced without helper virus. Gene Ther. 5: 938 –945. 23 Contag, C. H., et al. (1997). Visualizing gene expression in living mammals using a bioluminescent reporter. Photochem. Photobiol. 66: 523–531. 24 Contag, P. R., Olomu, I. N., Stevenson, D. K., and Contag, C. H. (1998). Bioluminescent indicators in living mammals. Nat. Med. 4: 245–247. 25 Chao, H., Mao, L., Bruce, A. T., and Walsh, C. E. (2000). Sustained expression of human factor VIII in mice using a parvovirus-based vector. Blood 95: 1594 –1599. 26 Brockstedt, D. G., Podsakoff, G. M., Fong, L., Kurtzman, G., Mueller-Ruchholtz, W., and Engleman, E. G. (1999). Induction of immunity to antigens expressed by recombinant adeno-associated virus depends on the route of administration. Clin. Immunol. 92: 67–75. 27 Lipshutz, G. S., Flebbe-Rehwaldt, L., and Gaensler, K. M. L. (2000). Re-expression following readministration of an adenoviral vector in adult mice after initial in utero adenoviral administration. Mol. Ther. 2: 374 –380. 28 Billingham, R. E., Brent, L., and Medawar, P. B. (1953). ‘Actively acquired tolerance’ of foreign cells. Nature 172: 603– 606. 29 Carrier, E., Lee, T. H., Busch, M. P., and Cowan, M. J. (1995). Induction of tolerance in

292

nondefective mice after in utero transplantation of major histocompatibility complexmismatched fetal hematopoietic stem cells. Blood 86: 4681– 4690. 30 Porada, C. D., et al. (1998). In utero gene therapy: Transfer and long-term expression of the bacterial neo(r) gene in sheep after direct injection of retroviral vectors into preimmune fetuses. Hum. Gene Ther. 9: 1571–1585. 31 Snyder, R. O., et al. (1997). Persistent and therapeutic concentrations of human factor IX in mice after hepatic gene transfer of recombinant AAV vectors. Nat. Genet. 16: 270 –276. 32 Miao, C. H., et al. (1998). The kinetics of rAAV integration in the liver. Nat. Genet. 19: 13–15. 33 Sweeney, T. J., et al. (1999). Visualizing the kinetics of tumor-cell clearance in living animals. Proc. Natl. Acad. Sci. USA 96: 12044 –12049. 34 Thompson, J. F., Hayes, L. S., and Lloyd, D. B. (1991). Modulation of firefly luciferase stability and impact on studies of gene regulation. Gene 103: 171–177. 35 DeLuca, M., and McElroy, W. D. (1974). Kinetics of the firefly luciferase catalyzed reactions. Biochemistry 13: 921–925. 36 Hall, J. C., Heel, K. A., Papadimitriou, J. M., and Platell, C. (1998). The pathobiology of peritonitis. Gastroenterology 114: 185–196. 37 Coutelle, C., Douar, A. M., Colledge, W. H., and Froster, U. (1995). The challenge of fetal gene therapy. Nat. Med. 1: 864 – 866. 38 Walsh, C. E. (1999). Fetal gene therapy. Gene Ther. 6: 1200 –1201. 39 Schneider, H., and Coutelle, C. (1999). In utero gene therapy: The case for. Nat. Med. 5: 256 –257. 40 de Wet, J. R., Wood, K. V., DeLuca, M., Helinski, D. R., and Subramani, S. (1987). Firefly luciferase gene: Structure and expression in mammalian cells. Mol. Cell. Biol. 7: 725–737. 41 Kim, D. W., Uetsuki, T., Kaziro, Y., Yamaguchi, N., and Sugano, S. (1990). Use of the human elongation factor 1 alpha promoter as a versatile and efficient expression system. Gene 91: 217–223. 42 Goodwin, E. C., and Rottman, F. M. (1992). The 3⬘-flanking sequence of the bovine growth hormone gene contains novel elements required for efficient and accurate polyadenylation. J. Biol. Chem. 267: 16330 –16334. 43 Lowry, O. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951). Protein measurement with the folin phenol reagent. J. Biol. Chem. 193: 265–275. 44 Osborn, M., and Isenberg, S. (1994). Cell Biology: A Laboratory Manual (J. Celis, Ed.), pp. 363–365. Academic Press, San Diego. 45 Thierry, A. R., et al. (1995). Systemic gene therapy: Biodistribution and long-term expression of a transgene in mice. Proc. Natl. Acad. Sci. USA 92: 9742–9746. 46 Glantz, S. A. (1997). Primer of Biostatistics, pp. 22, 26, 282–286. McGraw-Hill, New York.

MOLECULAR THERAPY Vol. 3, No. 3, March 2001 Copyright © The American Society of Gene Therapy