Efficient Transduction of Liver and Muscle after in Utero Injection of Lentiviral Vectors with Different Pseudotypes

In this study we investigate the efficacy of lentiviral vectors of different pseudotypes for gene transfer to tissues of the preimmune fetus. BALB/c fetuses at 14–15 days' gestation received lentiviral vectors carrying the transgene lacZ under the control of the human cytomegalovirus (CMV) promoter by intramuscular (i.m.) or intrahepatic (i.h.) injection. We pseudotyped the lentiviral vectors with vesicular stomatitis virus (VSV-G), with Mokola virus, or with Ebola virus envelope glycoproteins. We harvested the pups at time points between 5 days and 9 months following injection and performed a detailed histologic assessment. The efficiency and distribution of transduction after in utero administration was highly dependent upon the route of administration and the pseudotype of vector used. Biodistribution studies showed widespread distribution of vector sequences in multiple tissues, albeit at very low levels, and transduced cells were found in significant numbers only in liver, heart, and muscle. Overall, VSV-G was the most efficient in transducing hepatocytes, whereas Mokola and Ebola were more efficient in transducing myocytes. Transduction of cardiomyocytes was observed after both i.m. and i.h. injection of all three vectors. Our findings of long-term transduction of skeletal myocytes and cardiomyocytes after in utero administration suggest a novel strategy for the treatment of congenital muscular dystrophies. In this study we investigate the efficacy of lentiviral vectors of different pseudotypes for gene transfer to tissues of the preimmune fetus. BALB/c fetuses at 14–15 days' gestation received lentiviral vectors carrying the transgene lacZ under the control of the human cytomegalovirus (CMV) promoter by intramuscular (i.m.) or intrahepatic (i.h.) injection. We pseudotyped the lentiviral vectors with vesicular stomatitis virus (VSV-G), with Mokola virus, or with Ebola virus envelope glycoproteins. We harvested the pups at time points between 5 days and 9 months following injection and performed a detailed histologic assessment. The efficiency and distribution of transduction after in utero administration was highly dependent upon the route of administration and the pseudotype of vector used. Biodistribution studies showed widespread distribution of vector sequences in multiple tissues, albeit at very low levels, and transduced cells were found in significant numbers only in liver, heart, and muscle. Overall, VSV-G was the most efficient in transducing hepatocytes, whereas Mokola and Ebola were more efficient in transducing myocytes. Transduction of cardiomyocytes was observed after both i.m. and i.h. injection of all three vectors. Our findings of long-term transduction of skeletal myocytes and cardiomyocytes after in utero administration suggest a novel strategy for the treatment of congenital muscular dystrophies. IntroductionAdvances in prenatal screening and molecular diagnosis, combined with the impetus from the human genome project, make it likely that in the near future most genetic diseases will be diagnosed early in gestation. The prenatal diagnosis of genetic diseases will allow increasing opportunities to consider in utero gene therapy, particularly if advantages of prenatal treatment can be demonstrated over existing postnatal therapies.There are a number of potential advantages to prenatal gene therapy [1Yang E.Y. Flake A.W. Adzick N.S. Prospects for fetal gene therapy.Semin. Perinatol. 1999; 23: 524-534Abstract Full Text PDF PubMed Scopus (17) Google Scholar]. The small size of the fetus compared with the adult recipient provides a stoichiometric advantage with respect to vector to target cell ratio, while favoring increased compartmental and hematogenous distribution of the vector. The early gestational environment is also uniquely proliferative and contains an increased frequency of stem cells that migrate to and seed developing organs. Thus, transduction of stem cells with their subsequent migration and expansion during the normal process of ontogeny may result in an amplification of therapeutic effect. Finally, the early gestational fetus is immunologically naïve, allowing the potential for development of tolerance to the viral vector or transgene-encoded proteins. This could be a significant advantage, because immune reactions have hampered postnatal clinical and experimental gene therapy efforts [2Herzog R.W. High K.A. Problems and prospects in gene therapy for hemophilia.Curr. Opin. Hematol. 1998; 5: 321-326Crossref PubMed Scopus (18) Google Scholar].The main challenge for successful future prenatal gene therapy, as for postnatal gene therapy, is the development of safe and effective gene transfer techniques that allow long-term, tissue-specific, and regulated expression of the desired protein. Although the ideal gene transfer approach has not yet been developed, we [3Yang E.Y. Persistent postnatal transgene expression in both muscle and liver after fetal injection of recombinant adenovirus.J. Pediatr. Surg. 1999; 34: 766-772Abstract Full Text PDF PubMed Scopus (30) Google Scholar, 4Yang E.Y. Cass D.L. Sylvester K.G. Wilson J.M. Adzick N.S. Fetal gene therapy: efficacy, toxicity, and immunologic effects of early gestation recombinant adenovirus.J. Pediatr. Surg. 1999; 34: 235-241Abstract Full Text PDF PubMed Scopus (38) Google Scholar] and others [5Tran N.D. In utero transfer and expression of exogenous genes in sheep.Exp. Hematol. 2000; 28: 17-30Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar, 6Tran N.D. Induction of stable prenatal tolerance to β-galactosidase by in utero gene transfer into preimmune sheep fetuses.Blood. 2001; 97: 3417-3423Crossref PubMed Scopus (62) Google Scholar, 7Porada C.D. 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. 1998; 9: 1571-1585Crossref PubMed Scopus (108) Google Scholar, 8Porada C.D. Tran N.D. Zhao Y. Anderson W.F. Zanjani E.D. Neonatal gene therapy. Transfer and expression of exogenous genes in neonatal sheep following direct injection of retroviral vectors into the bone marrow space.Exp. Hematol. 2000; 28: 642-650Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar, 9Themis M. Successful expression of β-galactosidase and factor IX transgenes in fetal and neonatal sheep after ultrasound-guided percutaneous adenovirus vector administration into the umbilical vein.Gene Ther. 1999; 6: 1239-1248Crossref PubMed Scopus (64) Google Scholar, 10Mitchell M. Jerebtsova M. Batshaw M.L. Newman K. Ye X. Long-term gene transfer to mouse fetuses with recombinant adenovirus and adeno-associated virus (AAV) vectors.Gene Ther. 2000; 7: 1986-1992Crossref PubMed Scopus (54) Google Scholar, 11Sekhon H.S. Larson J.E. In utero gene transfer into the pulmonary epithelium.Nat. Med. 1995; 1: 1201-1203Crossref PubMed Scopus (92) Google Scholar, 12Larson J.E. Morrow S.L. Happel L. Sharp J.F. Cohen J.C. Reversal of cystic fibrosis phenotype in mice by gene therapy in utero.Lancet. 1997; 349: 619-620Abstract Full Text Full Text PDF PubMed Scopus (91) Google Scholar, 13Lipshutz, G. S., Flebbe-Rehwaldt, L., and Gaensler, K. M.2000. Reexpression following readministration of an adenoviral vector in adult mice after initial in utero adenoviral administration. Mol. Ther.2: 374–380,Google Scholar, 14Schneider H. Sustained delivery of therapeutic concentrations of human clotting factor IX — a comparison of adenoviral and AAV vectors administered in utero.J. Gene. Med. 2002; 4: 46-53Crossref PubMed Scopus (54) Google Scholar, 15Tarantal, A. F., et al.2001. Rhesus monkey model for fetal gene transfer: studies with retroviral-based vector systems. Mol. Ther.3: 128–138,Google Scholar] have demonstrated relatively efficient gene transfer to fetal tissues with in utero administration. We have developed a mouse model for analysis of fetal gene therapy and have previously shown that injection of adenoviral vector into preimmune fetal mice leads to high efficiency, long-term expression of lacZ in both muscle and liver, without evidence of immune rejection of transduced cells [3Yang E.Y. Persistent postnatal transgene expression in both muscle and liver after fetal injection of recombinant adenovirus.J. Pediatr. Surg. 1999; 34: 766-772Abstract Full Text PDF PubMed Scopus (30) Google Scholar] (S.B. and A.W.F., manuscript submitted)In this study, we investigate the safety, distribution, and long-term expression of lentiviral vectors of varying pseudotype after in utero injection in preimmune mouse fetuses. We chose lentiviruses because of their proven ability for integrational transduction of stem cells. Lentiviruses also may be altered by changing the vector pseudotype, which may allow preferential transduction of specific cell types. We report that lentiviral vectors provide high-efficiency, long-term gene expression in the murine fetus with the distribution of transduction dependent upon the mode of administration and the vector pseudotype.ResultsLentivirus Injection in Utero Leads to Slight Toxicity with VSV-G but Not with Mokola or Ebola PseudotypesTo determine whether there was inherent toxicity from the viral preparations or from early overexpression of the transgene, we compared survival after intrahepatic (i.h.) injection at 14–15 days’ gestation with each vector pseudotype with that after injection of phosphate-buffered saline (PBS) alone. Results are analyzed as the number of surviving pups at birth over the total number of pups injected for each group, expressed as a percentage (Table 1). There was a decrease in survival with the vesicular stomatitis virus (VSV-G) pseudotype at both titers tested compared with PBS, statistically significant at a titer of 1 × 1010/ml. There were no significant differences in survival with the Ebola or Mokola pseudotypes. Because the VSV-G pseudotyped vector was made in a different laboratory from the Mokola and Ebola pseudotypes, there may also be differences in vector preparation that lead to increased toxicity. However, the fact that survival is increased at the lower titer of VSV-G pseudotyped vector suggests that toxicity may be due to the high titer achieved with this vector rather than to factors intrinsic to the envelope glycoprotein. Additionally, at high titer, there may be free VSV-G that can be toxic. Finally, although numerous litters were included in each group, other factors, such as maternal cannibalism, may affect survival data.TABLE 1Survival after i.h. injection of virus compared with injection of PBSPBSVSV-GVSV-GMokolaEbola1 × 1010/ml1 × 109/ml2 × 108/ml1–3 × 108/mlNo. injected861181013041No. surviving5144471725% survival5937475761CHITEST value0.00000100.080.710.83 Open table in a new tab Biodistribution of VirusTo determine the extent of systemic spread to other tissues, we performed quantitative real-time PCR on multiple tissues of animals injected with the different vectors harvested either as neonates (1 week after injection; Table 2) or as adults (5–8 months after injection; Table 3). Spiking experiments showed that copies of viral sequences could be accurately detected with this technique (Table 4). We chose to analyze i.h. injected animals because this represents the equivalent of a systemic injection. At both time points, viral sequences were observed in tissues away from the site of injection, albeit at low levels. This is consistent with our histology data in that we have seen gene expression primarily in the liver and heart after i.h. injection and primarily in the muscle and heart after i.m. injection (a few positive cells have also been seen in the lung with all pseudotypes; data not shown). Earlier time-point animals showed higher amounts of vector sequence in some organs, especially the liver, compared with late time-point animals. This may be secondary to detection of preintegration complexes at the early time point, to transduction of cells of limited lifespan with subsequent cell death, and/or to dilution of transduced cells secondary to the exponential growth of the liver during the life of the animal. All animals analyzed demonstrated PCR evidence of vector sequences in one or more tissues.TABLE 2Biodistribution of lentiviral vectors 1 week after in utero i.h. injectionMouseLiverSpleenThymusHeartLungMuscleKidneyIntestineVSV-G1100nd0650.6ndnd0.423000000.90000358,000nd0.103.43.30.215,000Mokola40.600000.800528000.2283.11.710.4Ebola6250nd019017392177550713.125059743.10.4Number of vector copies detected by quantitative PCR in multiple tissues of individual mice 1 week after i.h. injection of VSV-G, Mokola, or Ebola pseudotyped vectors. nd, Not determined. Open table in a new tab TABLE 3Biodistribution of lentiviral vectors 5–8 months after in utero i.h. injectionMouseLiverSpleenThyHeartLungMscKidneyIntestBMBloodGonadsVSV-G820.11100.3000.10.40900.1000.30000001000000.200005.80110.31000000000.40Mokola12003.401400009013000000000.32.701411002.60.12.50000.90150.160000.70.3000352.2Ebola160.6300028.50000.8017180.10582.44.63.30.720.20.80185.900.201.80.76.6000001910006.51.610000.20Number of vector copies detected by quantitative PCR in multiple tissues of individual mice 5–8 months after in utero i.h. injection of VSV-G, Mokola, or Ebola pseudotyped vectors. Thy, thymus. Msc, muscle. Intest, intestine. BM, bone marrow. Open table in a new tab TABLE 4Spiking experiment for quantitative PCRCopy no. added to reactionCopy no. by PCRNegative control01 copy/100 ng0.510 copies/100 ng14100 copies/100 ng711000 copies/100 ng8401 × 104 copies/100 ng8.8 × 1031 × 105 copies/100 ng9.9 × 1041 × 106 copies/100 ng1.2 × 1061 × 107 copies/100 ng1.2 × 1071 × 108 copies/100 ng7.6 × 107We obtained copy numbers by spiking known Copies of positive control plasmid diluted into negative control mouse liver DNA. 100 ng = ∼ 15,000 cells. Open table in a new tab One animal that was injected with the Mokola vector had PCR-detectable sequences in the ovaries. Because of the small size of the mouse ovary, we used the entire ovary for DNA extraction, and adjacent frozen sections were not available for analysis. Although germline transduction has not been observed after in utero injection of integrating viruses [5Tran N.D. In utero transfer and expression of exogenous genes in sheep.Exp. Hematol. 2000; 28: 17-30Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar, 16Lipshutz, G. S., et al.2001. In utero delivery of adeno-associated viral vectors: intraperitoneal gene transfer produces long-term expression. Mol. Ther.3: 284–292,Google Scholar], the magnitude of the risk needs to be determined for each specific pseudotype and gestational age of injection in a study designed specifically to answer this question. Further refinement of the concept of pseudotyping with hybrid envelopes may assist in delivering viral vectors with specificity for individual tissues.We next wanted to test the relationship between the PCR data and the tissue histology. We chose to analyze hearts because we had seen expression in the heart at all time points for all pseudotypes and to minimize the possibility of sampling error by analyzing a small organ. In 20 animals, including those reported above, we used one half of the heart for genomic DNA extraction and the other half for analysis of frozen sections and scored levels of expression using a semi-quantitative scoring system. As expected, hearts that had a greater number of blue cells had higher vector copy numbers by PCR (Fig. 2). This experiment appears to validate the semi-quantitative scoring system that we have used in other organs.Intrahepatic Injection of Lentivirus Leads to Expression Primarily in the Liver, with Greatest Efficiency from VSV-G PseudotypeShort-term analysis of frozen sections of whole newborn pups after in utero i.h. injection demonstrated expression of marker gene primarily in the liver and in the heart, consistent with the results of the biodistribution study. Frozen sections of whole newborn pups 1 week after injection were uniformly negative in the thymus, spleen, and brain, but rare isolated cells were seen in the lung, diaphragm, and abdominal wall muscles near the site of injection with all pseudotypes (data not shown). Endogenous β-galactosidase staining in intestines and spleen precluded further analysis of these tissues.Transduction of liver cells appeared to occur in a uniform distribution throughout the liver, rather than remaining confined to the site of injection, reducing the likelihood of sampling error. For each pseudotype, animals were analyzed as neonates (5–10 days after injection) and at 4 weeks. Although some animals in each group were analyzed at later time points, up to 9 months, these livers were negative in all groups, except a focal area of the liver in one animal. Therefore, the semi-quantitative analysis of expression was performed only on livers at the 1- and 4-week time points. The cytomegalovirus (CMV) promoter is a weak promoter in the liver and may be lost at later time points. Silencing of the CMV promoter has been observed in the liver [17Loser P. Jennings G.S. Strauss M. Sandig V. Reactivation of the previously silenced cytomegalovirus major immediate-early promoter in the mouse liver: involvement of NFκB.J. Virol. 1998; 72: 180-190PubMed Google Scholar] but has not been described with lentivirus.In utero injection of all pseudotypes led to expression of lacZ in foci of hepatocytes in the liver (Figs. 1A–1C). High-power analysis of 2-mm plastic sections demonstrated that the transduced cells are octagonal cells with large nuclei and morphologically appear to be hepatocytes (Fig. 1D, VSV-G shown). At 10 days, these cells appeared to be in clusters, suggestive of clonal expansion of one transduced cell (Figs. 1E–1F). This pattern of expression persisted 4 weeks after injection in some animals (Fig. 1G, Mokola shown). Later time points failed to show expression, except in an isolated area of the liver of one animal (Fig. 1H). We observed gene expression by X-Gal staining only in hepatocytes, as defined by morphologic criteria.FIG. 1Histology of liver sections from mice following in utero i.h. injection of viral vector. All are frozen sections except (D) and (F), which are plastic. (A) VSV-G, 1.5 weeks, ×5. (B) Mokola, 1.5 weeks, ×5. (C) Ebola, 1.5 weeks, ×10. (D) VSV-G, 5 days, ×40. (E) VSV-G, 1.5 weeks, ×40. (F) Mokola, 4 weeks, ×40. (G) Mokola, 4 weeks, ×5. (H) Mokola, 5 months, ×40. (I) Negative control, ×20.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Semi-quantitative comparison of expression levels in the liver following i.h. injection is depicted in Fig. 3. As indicated, the Ebola pseudotype is significantly less efficient than those of VSV-G or Mokola in transducing hepatocytes, whereas there is no significant difference between the VSV-G and Mokola groups. Overall, we saw gene expression in liver in eight of nine VSV-G-, three of eight Mokola-, and two of nine Ebola-injected animals. The fact that there are many animals in the Mokola and Ebola groups that do not have gene expression in the liver also suggests differing tissue tropism for different vector pseudotypes. Many of these animals were positive in the heart, arguing against a technical error.FIG. 3Histology of plastic muscle sections from mice following in utero intramuscular injection of viral vector. (A) VSVG, 4 weeks, ×40. (B) Mokola, 10 weeks, ×20 (hamstring). (C) Ebola, 4 weeks, ×20. (D) Mokola, 10 weeks, ×20 (quadriceps). (E) Mokola, 6 months, ×20. (F) Ebola, 10 weeks, ×20. (G) Mokola, 6 months, ×20 (diaphragm). (H) Ebola, 10 weeks, ×20 (diaphragm). (I) Negative control, ×20.View Large Image Figure ViewerDownload Hi-res image Download (PPT)When we performed PCR analysis for vector sequences, there were surprisingly high copy numbers in the livers of animals injected with the Ebola vector. However, blinded scoring of livers from animals for which we had PCR data confirmed that there was no gene expression in these animals, even at early time points (when gene expression was seen with the other two pseudotypes). This suggests that the Ebola pseudotyped vector may transduce cell types in which the CMV promoter is very inefficient, leading to false-negative results with X-Gal immunohistochemistry.At the time of our in utero injection, hematopoiesis was confined to the fetal liver. It might therefore be expected that the vector would also transduce hematopoietic stem cells after i.h. injection, as has been seen after in utero injection of early gestational lamb [5Tran N.D. In utero transfer and expression of exogenous genes in sheep.Exp. Hematol. 2000; 28: 17-30Abstract Full Text Full Text PDF PubMed Scopus (51) Google Scholar, 7Porada C.D. 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. 1998; 9: 1571-1585Crossref PubMed Scopus (108) Google Scholar] and rat fetuses [18Clapp D.W. Dumenco L.L. Hatzoglou M. Gerson S.L. Fetal liver hematopoietic stem cells as a target for in utero retroviral gene transfer.Blood. 1991; 78: 1132-1139Crossref PubMed Google Scholar, 19Clapp D.W. Freie B. Lee W.H. Zhang Y.Y. Molecular evidence that in situ-transduced fetal liver hematopoietic stem/progenitor cells give rise to medullary hematopoiesis in adult rats.Blood. 1995; 86: 2113-2122PubMed Google Scholar]. In particular, VSV-G pseudotyped lentivirus has been shown to transduce hematopoietic stem cells (HSCs) after in utero injection of primate fetuses [15Tarantal, A. F., et al.2001. Rhesus monkey model for fetal gene transfer: studies with retroviral-based vector systems. Mol. Ther.3: 128–138,Google Scholar]. To determine whether HSCs can be transduced at the time of in utero injection, we injected some animals with VSV-G-lenti-CMV-GFP, which encodes green fluorescent protein (GFP) driven by the CMV promoter. In this group of animals, we analyzed peripheral blood and bone marrow (BM) by fluorescence-activated cell sorting (FACS) for GFP expression at 11 weeks. We detected no expression by FACS in eight of eight animals, although we observed gene expression in the liver (data not shown). PCR of blood and BM of these animals showed that seven of eight had vector sequences in blood and two of eight had sequences in BM (Table 5). To determine if the negative FACS results were due to poor expression from the CMV promoter, we injected a second group of animals with VSV-G-lenti-CAG-GFP, which encodes GFP driven by the CMV enhancer with the chicken β-actin promoter. Again, we detected no gene expression by FACS in the blood of these animals, but three of four were positive in blood and one of four was positive in BM by PCR (animals 8–11, Fig. 1B). The fact that BM was negative in many animals that were PCR-positive in the blood indicates that HSCs were not transduced or were transduced in extremely small numbers. Given the extremely low frequency of transduction as determined by quantitative PCR (copy numbers are calculated per 100 ng of tissue, about 15,000 cells), colony assays were not done in animals injected with this virus. Future experiments using a virus that includes a drug resistance gene to select for transduced cells, as well as testing of other lentiviral pseudotypes in this system, are planned to address this question.TABLE 5Quantitative PCR in liver, blood, and BM in animals after i.h. injection of VSV-G-CMV-GFPMouse numberLiverBloodBM98471.50.709848120.9098490.30098500.92.21.39853170.509854621.009855130.809856251.60.2Number of vector sequences detected in liver, blood, and BM specimens from animals injected with lenti-VSV-G-CMV-GFP intrahepatically, harvested at 11 weeks. These animals were negative for GFP expression by FACS analysis of blood and BM. Open table in a new tab Intramuscular Injection of Lentivirus Leads to Expression Primarily in the Injected Muscle Groups, with Greatest Efficiency from the Mokola and Ebola PseudotypesShort-term analysis demonstrated that in utero intramuscular injection of virus leads to expression primarily in the muscle groups in the injected limb, as well as in the heart. We therefore concentrated on these organs for further analysis. Frozen sections of whole newborn pups 1 week after injection were negative in the thymus, spleen, and brain, but rare isolated positive cells were seen in the liver with the VSV-G pseudotype. For each pseudotype, animals were analyzed at 4 weeks, 10–13 weeks, and 6 months.Because the left hindlimb is a very small target for in utero injections at 14 days’ gestation, frozen sections of muscle harvested from the left leg of adult animals can easily miss the site of injection and result in sampling error. We therefore stained the muscles immediately after harvesting with X-Gal to identify the positive areas for subsequent plastic section analysis. This method allowed us to determine the best score for each animal. In utero injection of all three pseudotypes led to transduction of myocytes, with much greater efficiency with the Mokola and Ebola pseudotypes compared with that of VSV-G (Figs. 3A–3C). In each animal, muscles from different compartments (quadriceps, hamstrings, back and gluteal muscles, and gastrocnemius muscles) were harvested. Many animals in the Mokola and Ebola groups were positive in more than one compartment, indicating broad distribution of virus away from the site of injection (FIG. 3, FIG. 4, Mokola shown). Expression persisted at 10 weeks (Fig. 3F, Ebola shown) and at 6 months at the latest time point tested (Fig. 3E, Mokola shown). Both Mokola and Ebola pseudotyped vectors also transduced myocytes in the diaphragm after injection in the left hindlimb (Figs. 3G and 3H).FIG. 4Histology of plastic and frozen sections of hearts after in utero i.m. or i.h. injection of viral vector. All are frozen sections except (D, G, H), which are plastic. (A) VSV-G, i.h. injection, 4 weeks, ×10. (B) Mokola, i.h. injection, 1.5 weeks, ×5. (C) Ebola, i.h. injection, 4 weeks, ×10. (D) Ebola, i.h. injection, 9 weeks, ×100. (E) Ebola, i.m. injection, 10 weeks, ×5. (F) Mokola, i.h. injection, 6 months, ×40. (G) Ebola, i.h. injection, 4 weeks, ×100. (H) Negative control heart, ×20.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Semi-quantitative scoring of the muscles from the animals at the 4 week, 10–13 week, and 6 month time points is shown in Fig. 2C. As shown, VSV-G is significantly less efficient than Mokola or Ebola in transducing myocytes; there is no difference between the Mokola and Ebola groups. Overall, 3 of 13 VSV-G-, 9 of 13 Mokola-, and 9 of 12 Ebola-injected animals had evidence of transgene expression. We saw robust gene expression in the Mokola group 6 months following injection but not in the Ebola group at the same time point; we interpret this as interanimal variability, because the same promoter was used in all pseudotypes.In Utero Injection of All Pseudotypes Results in Systemic Distribution, with Transduction of Cardiomyocytes after Both i.h. and i.m. InjectionWe observed transduction of cardiomyocytes after injection of all three pseudotypes, indicating systemic spread of the virus. Predictably, the efficiency of cardiomyocyte transduction mirrored that of skeletal muscle, with superior results from the Mokola and Ebola pseudotypes compared with VSV-G (Figs. 4A–4C). For both the Mokola and Ebola pseudotypes, a single in utero injection led to transduction of cardiomyocytes throughout the heart (Figs. 4B and 4C, respectively). High-resolution plastic sectioning confirmed that the transduced cells have the morphology of cardiomyocytes (Fig. 4D). In some animals, significant numbers of transduced cardiomyocytes were seen along the lumen of the ventricle, supporting hematogenous spread and increased transduction of the vascular endocardial region (Fig. 4E). Expression was long-term, up to 6 months at the last time point tested (Fig. 4F). Both i.m. and i.h. injection led to transduction of cardiac myocytes. In addition to cardiomyocytes, we observed transduction of cells morphologically consistent with capillary endothelial cells with the Ebola pseudotype (Fig. 4G).DiscussionLentiviruses have the ability to transduce a wide variety of cells, including nondividing cells, and to integrate into the genome to provide sustained gene expression. Recently, it has been demonstrated that changing the pseudotype of lentiviruses can modify their tissue tropism [20Kobinger G.P. Weiner D.J. Yu Q.C. Wilson J.M. Filovirus-pseudotyped lentiviral vector can efficiently and stably transduce airway epithelia in vivo.Nat. Biotechnol. 2001; 19: 225-230Crossref PubMed Scopus (264) Google Scholar, 21Mochizuki H. Schwartz J.P. Tanaka K. Brady R.O. Reiser J. High-titer human immunodeficiency virus type 1-based vector systems for gene delivery into nondividing cells.J. Virol. 1998; 72: 8873-8883Crossref PubMed Google Scholar]. The VSV-G envelope has been extensively used because it can transduce a variety of cell types and because it is stable, allowing concentration of the vector by