Transgenic Monkeys Produced by Retroviral Gene Transfer into Mature Oocytes

See allHide authors and affiliations

Science  12 Jan 2001:
Vol. 291, Issue 5502, pp. 309-312
DOI: 10.1126/science.291.5502.309


Transgenic rhesus monkeys carrying the green fluorescent protein (GFP) gene were produced by injecting pseudotyped replication-defective retroviral vector into the perivitelline space of 224 mature rhesus oocytes, later fertilized by intracytoplasmic sperm injection. Of the three males born from 20 embryo transfers, one was transgenic when accessible tissues were assayed for transgene DNA and messenger RNA. All tissues that were studied from a fraternal set of twins, miscarried at 73 days, carried the transgene, as confirmed by Southern analyses, and the GFP transgene reporter was detected by both direct and indirect fluorescence imaging.

Although transgenic mice have been invaluable in accelerating the advancement of biomedical sciences (1–5), many differences between humans and rodents have limited their usefulness (6–9). The major obstacle in producing transgenic nonhuman primates has been the low efficiency of conventional gene transfer protocols. By adapting a pseudotyped vector system, efficient at up to 100% in cattle (10, 11), we circumvented problems in traditional gene transfer methodology to produce transgenic primates.

We injected 224 mature rhesus oocytes with high titer [108to 109 colony-forming units (cfu)/ml] moloney retroviral vector pseudotyped with vesicular stomatitis virus envelope glycoprotein G (VSV-G pseudotype) into the perivitelline space (Fig. 1; Table 1; 1012). The VSV-G pseudotype carried the GFP gene under the control of either the cytomegalovirus early promoter (CMV) [referred to as LNCEGFP-(VSV-G)] or the human elongation factor-1 alpha promoter (hEF-1α) [referred to as LNEFEGFP-(VSV-G)] (13). Because ∼10 to 100 pl was introduced into the perivitelline space, between 1 and 10 vector particles were introduced using LNCEGFP-(VSV-G) [109 cfu/ml] and between 0.1 to 1 with LNEFEGFP-(VSV-G) (108 cfu/ml). Oocytes were cultured for 6 hours before fertilization by intracytoplasmic sperm injection (ICSI). Vector particles incorporated into the oocyte in <4.5 hours as imaged by electron microscopy (14). Fifty-seven percent (n = 126) of embryos developed beyond the four-cell stage and 40 embryos were transferred to 20 surrogates, each carrying two embryos (Table 1). Rates for reproductive parameters are: fertilization [77% ICSI controls (15) versus 75% transgenesis], embryonic development [75% ICSI controls (15) versus 57% transgenesis], and implantation [66% ICSI controls (16) versus 25% transgenesis]. Most control ICSI pregnancies result in live offspring (83%) (16).

Figure 1

Injection of VSV-G pseudotyped retroviral vector, enclosing the GFP gene and protein, into the perivitelline space of mature rhesus oocytes. (A) Transmitted light and (B) epifluorescence imaging of GFP carried within the vector particles. Magnification: ×100.

Table 1

Transgenesis efficiency in rhesus embryos, fetuses, and offspring.

View this table:

Five pregnancies resulted in the births of three healthy males (Table 1, Fig. 2). A set of fraternal twins miscarried at 73 days (150 to 155 days normal gestation) and a blighted pregnancy (implantation attempt without a fetus) also occurred. One fetal twin of the miscarriage was an anatomically normal male, while the other was largely resorbed in utero. The three births and the blighted pregnancy resulted from nine embryo transfers in which LNEFEGFP-(VSV-G) was used, whereas the twin pregnancy was established from 11 embryo transfers with LNCEGFP-(VSV-G) (Table 1).

Figure 2

(A) Transgenic rhesus male with inserted DNA (“ANDi”). GFP expression was observed in hair shafts (B) and toenails (C) by direct epifluorescent examination in the male stillborn but not in the accessible tissues from ANDi. Immunostaining and epifluorescent examination of placental frozen sections from the male stillborn demonstrates the presence of the GFP protein. (D) Anti-GFP detection in placenta by rhodamine (red) immunofluorescent microscopy. (E) GFP detection by fluorescein (green) epifluorescence of the same section demonstrates the direct expression of the transgene. (F) Overlay of the green (E) and red (D) images demonstrates colocalization of direct GFP fluorescence with anti-GFP imaging. Blue, Hoechst 33342 DNA staining. Magnification in (D) through (F): ×400.

Transgene integration, transcription, and expression from the newborns were examined in hair, blood, umbilical cords, placentae, cultured lymphocytes, buccal epithelial cells, and urogenital cells passed in urine, along with 13 tissues from the male stillborn, nine from the resorbed one, and specimens from the blighted pregnancy (17). Polymerase chain reaction (PCR) was performed with primer sets that covered the flanking region of the vector pLNC-EGFP or pLNEF-EGFP and the GFP gene (18). One newborn, ANDi, showed the presence of the transgene in all analyzed tissues, and the transgene was present in all tissues analyzed from both stillbirths including placentae and testes (Fig. 3). Total RNA was extracted for standard reverse transcription followed by PCR amplification (RT-PCR) with primer sets specific for the transgene (18). Transgene transcription was demonstrated in all of the tissues in the fetuses and in the accessible tissues from the infant carrying the transgene (Fig. 3).

Figure 3

PCR and RT-PCR analyses of transgenic and control tissues. (A) Thirteen tissues from an intact fetus were submitted for PCR and (B) 11 tissues for RT-PCR. (C) Analysis of the male stillborn. Tissues from the reabsorbed fetus were collected from eight different regions to ensure broad representation, because precise anatomical specification was limited. (D through F) PCR, RT-PCR of the reabsorbed fetus. A total of seven samples were obtained from each offspring for PCR (G), two samples for RT-PCR (H) from “ANDi” and one of the other two male offspring. (I) Analysis of the newborns, indicates that “ANDi” is a transgenic male with the presence of mRNA in all analyzed tissues. Co, cord; Bo, blood; Ly, lymphocyte; Bu, buccal cells; Ur, urine; Ha, hair; Pl, placenta; Lu, lung; Li, liver; He, heart; In, intestine; Ki, kidney; Bl, bladder; Te, testis; Mu, muscle; Sk, skin; Ta, tail; Pa, pancreas; Sp, spleen; T1 = placenta from reabsorbed fetus; T2 to T9 = tissues retrieved from eight regions of the reabsorbed fetus; C1 = nontransgenic rhesus tissue; C2 = C1 + pLNC-EGFP; C3 = ddH2O; C4 = 293GP-LNCEGFP packaging cell; C5 = nontransgenic liver; C6 = transgenic lung without DNase; C7 = transgenic lung without reverse transcription; C8 = C1 + pLNEF-EGFP. ND, not determined.

Southern blot analysis of 10 tissues from the male stillbirth and eight samples from the other twin demonstrated multiple integration sites into their genomic DNA (Fig. 4) (19). Vector integration was determined by PCR of placenta, cord, blood, hair, and buccal cells using a primer set specific for the unique retroviral long terminal repeat (LTR) regions indicative of successful provirus integration into the host genome (20,21). This provirus sequence was found in one infant and both stillbirths (Fig. 4D). Infant welfare considerations limited tissue availability, and genomic DNA obtained was insufficient for Southern analysis. The male infant with the inserted transgene has been named “ANDi” (for “inserted DNA,” in a reverse transcribed direction;Fig. 2A).

Figure 4

(A) Southern blot analysis of Hind III (single digestion site) digested genomic DNA. Full-length GFP labeled with [32P] was used as a probe to detect the transgene, which was detected in genomic DNA of the normal male stillbirth (B) and reabsorbed fetus (C). Nontransgenic rhesus tissue was used as a negative control (C1) and pLNC-EGFP DNA as a positive control (not shown). Various sized fragments were demonstrated in tissues obtained from each. This result indicates multiple integration sites due to the use of a restriction enzyme with a single digestion site within the transgene. (D) Detection of the unique provirus sequence. A total of five tissues from each infant and two tissues from a male stillbirth and the reabsorbed fetus were submitted for PCR. Provirus sequence was detected in “ANDi” and the two stillbirths (42), which indicates that they are transgenic. Abbreviations are the same as those in Fig. 3. Mu, muscle from the male stillborn; T3, tissue from the reabsorbed fetus.

GFP direct fluorescence in the toenails and hair of the fetus, as well as the placenta (Fig. 2, B through F), provided further evidence of transgenesis. Colocalization between direct GFP fluorescence and indirect anti-GFP immunocytochemical imaging demonstrated that the GFP protein is found exclusively at the direct fluorescence sources (Fig. 2, D through F). Furthermore, neither direct fluorescein nor indirect rhodamine fluorescence was observed in controls (22). Because tissues from the fetus originated from the three germ layers, the timing of transgene integration may have occurred before implantation, perhaps even before the first DNA replication cycle (10). The high efficiency of this approach has been linked to the absence of the nuclear envelope in oocytes naturally arrested in second meiotic metaphase (10, 23).

The miscarriage is likely due to the twin pregnancy, which is rare and high-risk in rhesus. The twin stillbirth originated from the higher titer vector, whereas the three births, including the transgenic one, and the blighted pregnancy originated from the lower titer LNEFEGFP-(VSV-G) vector (108 cfu/ml; Table 1). Although only one live offspring is shown to be transgenic, we cannot yet exclude the possibility of transgenic mosaics in the others. We have neither demonstrated germline transmission nor the presence of transgenic sperm; this must await ANDi's development through puberty in about 4 years. Vector titers and volume injected may play crucial roles in gene transfer efficiency. These offspring and their surrogates are now housed in dedicated facilities with ongoing, stringent monitoring.

Nonhuman primates are invaluable models for advancing gene therapy treatments for diseases such as Parkinson's (24) and diabetes (25), as well as ideal models for testing cell therapies (26) and vaccines, including those for HIV (27, 28). Although we have demonstrated transgene introduction in rhesus monkeys, significant hurdles remain for the successful homologous recombination essential for gene targeting (29). The molecular approaches for making clones [either by embryo splitting (30) or nuclear transfer (31–36)], utilizing stem cells (37–39), and now producing transgenic monkeys, could be combined to produce the ideal models to accelerate discoveries and to bridge the scientific gap between transgenic mice and humans.

  • * To whom correspondence should be addressed. E-mail: schatten{at}


View Abstract

Navigate This Article