Adenovirus-mediated gene delivery into mousespermatogonial stem cellsMasanori Takehashi*, Mito Kanatsu-Shinohara*†, Kimiko Inoue‡, Narumi Ogonuki‡, Hiromi Miki‡, Shinya Toyokuni§,Atsuo Ogura‡, and Takashi Shinohara*¶

*Department of Molecular Genetics, †Horizontal Medical Research Organization, and §Department of Pathology and Biology of Diseases, Graduate Schoolof Medicine, Kyoto University, Kyoto 606-8501, Japan; and ‡Bioresource Center, Institute of Physical and Chemical Research (RIKEN), Ibaraki 305-0074, Japan

Edited by Ryuzo Yanagimachi, University of Hawaii, Honolulu, HI, and approved December 26, 2006 (received for review October 21, 2006)

Spermatogonial stem cells represent a self-renewing populationof spermatogonia, and continuous division of these cells sup-ports spermatogenesis throughout the life of adult male ani-mals. Previous attempts to introduce adenovirus vectors intospermatogenic cells, including spermatogonial stem cells, havefailed to yield evidence of infection, suggesting that male germcells may be resistant to adenovirus infection. In this study weshow the feasibility of transducing spermatogonial stem cells byadenovirus vectors. When testis cells from ROSA26 Cre reportermice were incubated in vitro with a Cre-expressing adenovirusvector, Cre-mediated recombination occurred at an efficiency of49 –76%, and the infected spermatogonial stem cells couldreinitiate spermatogenesis after transplantation into seminifer-ous tubules of infertile recipient testes. No evidence of germ-lineintegration of adenovirus vector could be found in offspringfrom infected stem cells that underwent Cre-mediated recom-bination, which suggests that the adenovirus vector infected thecells but did not stably integrate into the germ line. Neverthe-less, these results suggest that adenovirus may inadvertentlyintegrate into the patient’s germ line and indicate that there isno barrier to adenovirus infection in spermatogonial stem cells.

gene therapy � germ cell � spermatogenesis

Spermatogenesis depends on the continuous proliferation ofspermatogonial stem cells. Although the number of sper-

matogonial stem cells is very small (comprising only 0.2–0.3% inthe mouse testis) (1, 2), these cells undergo self-renewingdivision and produce differentiated cells while maintaining anundifferentiated state. Because spermatogonial stem cells trans-mit genetic information to the offspring, the introduction ofgenetic material into spermatogonial stem cells results in per-manent modification of the germ line.

However, initial attempts to introduce genetic material intospermatogonial stem cells met with little success in part becauseof a lack of methods for transducing genetic material into stemcells. The first reported evidence of germ-line transduction usedretrovirus vectors, which have relatively high infection efficiencyand have been widely used in the transduction of stem cells inseveral self-renewing tissues (3). Spermatogonial stem cells wereinfected with retrovirus vectors in vitro and transplanted into theseminiferous tubules for offspring production. Transplantedstem cells colonized the empty seminiferous tubules of infertilerecipient testes and reinitiated spermatogenesis, eventually lead-ing to the production of transgenic animals (4, 5). Although theseresults opened up an opportunity for in vitro genetic manipula-tion of spermatogonial stem cells, they revealed a serious safetyconcern in human somatic gene therapy. Inadvertent infection ofgerm-line cells by gene therapy vectors could lead to verticalgerm-line transmission of the virus and potential insertionalmutagenesis.

Adenovirus is another important type of viral vector that isused in human gene therapy (6). Unlike retrovirus vectors, whichcan infect only dividing cells, adenovirus has relatively hightransduction efficiency in target cells and infects both dividing

and nondividing cells (6). Although adenovirus vectors can beprepared at higher titer than retrovirus vectors and infect a largerange of host cells, including hematopoietic stem cells or em-bryonic stem cells (7, 8), many in vitro and in vivo attempts totransduce spermatogenic cells have not provided evidence ofinfection. In particular, the absence of a functional assay todirectly measure stem cell activity has interfered with a conclu-sive determination of whether spermatogonial stem cells can beinfected by adenoviruses. Although viruses that are injectedintravenously or into nongonadal tissues can reach mouse orhuman testes, only Sertoli cells are preferentially infected byintratesticular injection, and there is no evidence of germ cellinfection (9–14). Furthermore, there is no sign of infection afterdirect in vitro exposure of spermatogenic cells or mature spermto adenovirus (15, 16). These studies suggest that male germcells, including spermatogonial stem cells, cannot be infected byadenovirus vectors.

The purpose of this study is to examine the feasibility ofadenovirus-mediated gene delivery to spermatogonial stemcells. We hypothesized that adenovirus infection did not occur inprevious studies because of inefficient exposure of germ cells (orstem cells) to adenovirus and/or low sensitivity of the detectionmethods. To overcome these problems, we took advantage of theROSA26 reporter mouse strain, which can sensitively monitorCre-mediated deletion (17). By adding Cre recombinase intocells, it is possible to excise loxP-f lanked DNA sequences intransfected cells (17). An adenovirus expressing the Cre recom-binase gene was used to infect an enriched population ofspermatogonial stem cells in vitro, and infected cells weretransplanted into the seminiferous tubules of infertile animals.Recipient testes were analyzed for reporter gene expression, andoffspring DNA was examined for virus gene integration.

ResultsInfection of Spermatogonial Stem Cells in Vivo. In preliminaryexperiments we examined whether it is possible to infect spermato-gonial stem cells in vivo. An adenovirus vector that expresses LacZgene (AxCANLacZ) was introduced into the seminiferous tubulesof immature and mature wild-type mice. In contrast to i.v. delivery,microinjection into seminiferous tubules allows more efficientdirect exposure of germ cells to adenovirus. In particular, immaturetestes should provide better accessibility to spermatogonial stemcells because of the absence of a blood–testis barrier and multiplelayers of germ cells (18). Previous studies using immature testeshave shown efficient in vivo retrovirus infection of spermatogonial

Author contributions: M.K.-S. and T.S. designed research; M.T., M.K.-S., K.I., N.O., H.M., S.T.,A.O., and T.S. performed research; T.S. contributed new reagents/analytic tools; M.T.,M.K.-S., S.T., and T.S. analyzed data; and M.K.-S. and T.S. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS direct submission.

Abbreviations: GS cell, germ-line stem cell; mGS cell, multipotent GS cell.

¶To whom correspondence should be addressed. E-mail: tshinoha@virus.kyoto-u.ac.jp.

© 2007 by The National Academy of Sciences of the USA

2596–2601 � PNAS � February 20, 2007 � vol. 104 � no. 8 www.pnas.org�cgi�doi�10.1073�pnas.0609282104

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021

stem cells (19). However, when we analyzed the injected testes 4–8weeks after adenovirus injection we found only patchy infection inthe Sertoli cells that received adenovirus at the immature stage (Fig.1A). Although staining was also found in mature testes, this wassimilarly localized exclusively in the Sertoli cells, and we were notable to obtain clear evidence of germ cell infection (Fig. 1B). Testestransduced with adenovirus vector occasionally showed signs ofinflammation, but spermatogenesis and expression of the LacZgene were observed in all treated testes. In agreement with theseresults, when we microinjected EGFP-expressing adenovirus(AxCANEGFP) into mature testes, none of the 12 injected malessired transgenic progeny after mating with wild-type females.Although �100 offspring were analyzed for EGFP expression, wedid not observe EGFP fluorescence (data not shown).

Infection of Spermatogonial Stem Cells in Vitro. We next sought toexamine whether adenovirus can infect spermatogonial stem cellsin vitro. We dissociated immature testes and cultured them in thepresence of AxCANEGFP. The immature testes contain enrichedpopulations of spermatogonial stem cells because they lack differ-entiated cells (18). After infection, EGFP-expressing cells werefound on the next day in culture (Fig. 1 C and D). To determinewhether germ cells could be infected by the adenovirus, the infectedcells were analyzed by flow cytometry 2 days after infection. Theanalysis revealed that 13.3% of the EGFP-positive cells expressedthe spermatogonia marker EpCAM (Fig. 1I) (20).

To further examine whether the virus can infect spermatogo-nial stem cells, we exposed AxCANEGFP to cultured spermato-gonial stem cells or germ-line stem cells (GS cells) (21). Thesecells can undergo self-renewing division in the presence of glialcell line-derived neurotrophic factor in vitro, but they reinitiatespermatogenesis after transplantation into seminiferous tubules.After overnight infection with AxCANEGFP, most of the GScell colonies showed EGFP fluorescence (Fig. 1 E, F, and J). TheEGFP expression continued after several passages. Within 2–3weeks the fluorescence was barely detectable, which suggestedthat the vector did not integrate stably in the genome of the GScells (Fig. 1L). Although we did not find a significant effect ofvirus on the growth and morphology of GS cells at low virusconcentrations (6 � 103 pfu/ml), exposure to higher concentra-tions of adenovirus had a negative effect on GS cell growth, andonly 13% of the cultured cells could be recovered at 6 � 105

pfu/ml 6 days after infection (Fig. 1K). Adenovirus could alsoinfect ES-like multipotent GS cells (mGS cells) (Fig. 1 G and H)(22). These results established that spermatogonia can be in-fected in vitro by adenovirus vector.

Spermatogenesis from Adenovirus-Infected Stem Cells. To establishthat spermatogonial stem cells are infected by adenovirus, it isnecessary to test whether infected cells can initiate andmaintain long-term spermatogenesis by spermatogonial trans-plantation, which is the only functional assay for spermatogo-nial stem cells (23). However, we assumed that detecting signsof infection might be difficult in germ cell colonies becauseadenovirus generally remains as an episomal element and maydisappear as single stem cells undergo multiple rounds ofdivision to mature into spermatozoa (24). In agreement withthis possibility, the intensity of EGFP fluorescence in theAxCANEGFP-infected GS cells decreased during passages(Fig. 1L), and the injection of AxCANEGFP-infected testiscells into seminiferous tubules of infertile adult recipientanimals did not yield clear evidence of germ cell transduction(data not shown).

To establish a more sensitive method for detection of viralgene transduction, we took advantage of the Cre recombinasesystem (Fig. 2) (25). In this experiment, testis cells used forinfection were collected from ROSA26 Cre reporter (R26R)mice that were 5–10 days old. At this stage, spermatogonial stem

cells are enriched in the testis because of the absence ofdifferentiated cells. The testis cells were exposed overnight toCre-expressing adenovirus (AxCANCre) in vitro (26), and the

Fig. 1. Adenovirus-mediated gene transfer into male germ cells. (A and B)Histological appearance of testes injected with AxCANLacZ at 7 days (A) or 4weeks (B) after birth. Whole mounts of testes were stained 1 month after virusinjection. (C–H) In vitro infection of immature testis cells (C and D), GS cells (Eand F), and mGS cells (G and H) by AxCANEGFP. The cells were exposed toadenovirus overnight at 2.5 � 105 pfu/ml (C and D) or 1.8 � 105 pfu/ml (E–H),and EGFP fluorescence was examined 6 h (D) or 1 day (F and H) after infection.(I) Flow-cytometric analysis of immature testis cells 2 days after transductionof AxCANEGFP at 2.5 � 105 pfu/ml. EpCAM-positive spermatogonia cellsshowed EGFP fluorescence. Black line, control Ig; red line, specific antibody. (J)Flow-cytometric analysis of GS cells 3 days after transduction of AxCANEGFP.The values are mean � SEM (n � 3). (K) Increase in GS cell number afteradenovirus infection. After overnight infection, GS cells were cultured for 6days. The values are mean � SEM (n � 4). Although no significant differencein cell number was found at 6.0 � 103 pfu/ml, GS cell growth was inhibited athigher virus concentrations (P � 0.05 by t test). (L) The intensity of EGFP signalsdecreased during a 17-day culture. The cells were infected at 2.5 � 105 pfu/mland passaged twice during this period. (Scale bars: 50 �m for A and B and 100�m for C–H.)

Takehashi et al. PNAS � February 20, 2007 � vol. 104 � no. 8 � 2597

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021

infected cells were collected by trypsin digestion on the next dayfor transplantation into the seminiferous tubules of histocom-patible recipient mice that were 5–10 days old, allowing theefficient colonization of donor cells (18). Successful Cre-mediated recombination in infected cells would result in thedeletion of DNA sequences that are flanked by loxP sequences.Because ROSA26 promoter is active during spermatogenesis,the infected cells start to express LacZ gene after Cre-mediatedrecombination, which can be readily detected by X-Gal stain-ing (27).

Four separate experiments were performed, and a total of 17testes in 14 recipient animals were microinjected with thecultured cells. PCR and Southern blot analyses of the culturedcells revealed that 49–76% of the cells underwent Cre-mediatedrecombination after overnight incubation with Cre-expressingadenovirus (Fig. 3). This deletion occurred only when the cellswere exposed to AxCANCre (Fig. 3B), and sequence analysis ofthe cultured cells confirmed that loxP regions were maintainedafter culture without the presence of AxCANCre (data notshown). Three months after transplantation, testes were recov-ered from some of the recipients and stained for LacZ activity.LacZ-expressing colonies were found in all recipient testes,indicating that Cre recombinase was successfully expressed andinduced recombination in spermatogonial stem cells (Fig. 4A).Histological analysis of the recipient testes showed normal-appearing spermatogenesis, and mature spermatozoa could befound in the germ cell colony (Fig. 4B). LacZ-expressing roundand elongated spermatids were also observed. Because sper-matogonial stem cells are the only cell type that can establishcomplete spermatogenesis after transplantation, these resultsshow that spermatogonial stem cells infected with adenovirusvector could induce normal spermatogenesis.

Lack of Adenovirus Integration in Offspring from Infected Stem Cells.To determine whether the germ cells from the infected cells arefertile, we used in vitro microinsemination, a technique com-monly used to produce offspring in animals and humans (28, 29).Testes or epididymides were collected from three differentrecipients 7 months after transplantation of adenovirus-infecteddonor cells. Seminiferous tubules or epididymis were dissectedby fine forceps and mechanically dissociated by repeated pipet-ting. The recovered round spermatids and elongated spermatidsand spermatozoa were microinjected into oocytes from C57BL/6(B6) � DBA/2 F1 (BDF1) females. After 24 h of culture, 23 of123 (19%) of the embryos developed to the two-cell stage, andthey were transferred into the oviducts of pseudopregnant ICRfemales. Of the 85 embryos transferred, 46 (54%) implanted inthe uteri and 21 offspring were born (Table 1). Overall, offspringwere obtained from three of four recipient males that were used

in microinsemination, and PCR and Southern blot analysesrevealed that successful recombination occurred in all 14 off-spring that carried ROSA floxed allele (Fig. 4 D and E). Nineoffspring, six males and three females, grew up to be normalfertile adults. When we examined the expression of the LacZgene, three of the nine offspring showed X-Gal staining in skin

Fig. 2. Diagram of the experimental procedure. Testis cells from donor R26R mice were dissociated by trypsin digestion and infected in vitro by AxCANCreadenovirus. Cre-mediated recombination removed the neo cassette, and LacZ gene expression was initiated under the ROSA26 ubiquitous promoter. Theinfected cells were transplanted into infertile recipient testes. At 20 weeks after transplantation, recipient testes were mechanically dissociated, andspermatogenic cells were microinjected into oocytes to produce offspring. DNA from the offspring was analyzed by Southern blotting and PCR for integrationof adenovirus.

Fig. 3. Deletion of floxed sequence in ROSA26 locus. (A) Diagram of theexperimental design to detect deletion of floxed sequence by PCR and South-ern blot analyses. (B) PCR analysis of the AxCANCre-mediated deletion by invitro infection of immature testis cells. Deletion was detected only when thecultured cells were exposed to AxCANCre. Deletion did not occur whenAxCANEGFP was used for infection. (C) Deletion efficiency of ROSA26 locus.Genomic DNA of infected cells was digested with EcoRV and hybridized witha ROSA26-specific probe (see Materials and Methods). Levels of percentagedeletion, estimated by the intensity of each band, are indicated below.

2598 � www.pnas.org�cgi�doi�10.1073�pnas.0609282104 Takehashi et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021

biopsy. The ROSA26 promoter is ubiquitously active (17), andLacZ expression was also found in many organs (brain, liver,kidney, and testis) of the offspring (Fig. 4C). Normal offspringwere produced from both males and females, and the parentalgenotype was transmitted in a Mendelian manner to subsequentgenerations after natural mating (Fig. 4F).

To determine whether the adenovirus DNA integrated intothe genome of the offspring, DNA was collected from placentasor tails of the offspring and analyzed by PCR and Southernblotting, using the probe from adenovirus genome. Viral DNAwas not detected in any of the 21 offspring using either method,indicating that the adenovirus vector did not integrate into thegerm line (Fig. 5).

DiscussionPrevious attempts failed to demonstrate the infection of sper-matogenic cells by adenovirus vectors. In our recent study we

found that introduction of adenovirus into the seminiferoustubules resulted in transduction of Sertoli cells, but there was noevidence of infection in germ cells, which are more abundant inthe testes (13). This occurred even though germ cells expressadenovirus receptors (30, 31). Likewise, other in vivo and in vitrostudies also reported that only somatic cells, but not germ cells,are infected (9–16). Nevertheless, the current study now dem-onstrates that spermatogonial stem cells are susceptible toadenovirus infection.

Several factors led to successful adenovirus infection in sper-matogonial stem cells. First, we used a sensitive reporter mousestrain for virus infection (17). Previous studies have relied onmarker gene expression from virus vector or DNA analysis.Although these methods successfully detected retrovirus infec-tion, we assumed that they might fail to detect transient infec-tions or small amounts of infective viruses. In contrast, deletionof floxed sequences in the host genome is irreversible and doesnot require the presence of adenovirus at the time of detection.Second, we used a chicken �-actin and cytomegalovirus en-hancer promoter that has been used to express exogenous genesin spermatogenic cells (32, 33). It is known that germ-line cellssuppress viral promoters, and the use of these in previous studiesmay be one reason that adenovirus infection was not detected(34). Third, we used a higher concentration of spermatogonialstem cells. Previous studies used adult testes, in which theconcentration of stem cells is significantly low (0.2–0.3%) (1, 2).In contrast, we used a single-cell suspension prepared fromimmature testis or GS cells, which lack differentiated germ cellsand are more enriched for spermatogonial stem cells, leading tohighly efficient transduction.

Fig. 4. LacZ expression after Cre recombination. (A) Macroscopic appearanceof a recipient testis that received Cre-infected testis cells 3 months aftertransplantation. Blue tubules represent colonization of donor stem cells thatwere infected by AxCANCre adenovirus. (B) Histological appearance of arecipient testis. Note complete spermatogenesis. LacZ expression was found inall spermatogenic cells. (C) Macroscopic appearance of X-Gal-stained brain,kidney, testis, and liver of F1 offspring from AxCANCre-infected spermatogo-nial stem cells. Note ubiquitous LacZ expression. (D) Southern blot analysis ofEcoRV-digested tail genomic DNA from mature F1 offspring hybridized with aROSA26-specific probe. Two offspring showed Cre-mediated recombination.(E) PCR analysis of the deletion. The deletion of floxed sequence was con-firmed by PCR detection method, as shown in Fig. 3A. (F) Pedigree of an F1

male demonstrating transmission of parental genotype for three generations.Solid symbols indicate LacZ expression in these progeny. (Scale bars: 1 mm forA and 20 �m for B.)

Table 1. In vitro microinsemination with spermatogenic cellsrecovered from recipient W mice

Type of cellsinjected

No. of embryostransferred*

No. of embryosimplanted (%)

No. ofpups (%)

Round spermatid 19 10 (53) 0 (0)Elongated spermatid 6 4 (67) 0 (0)Testicular sperm 50 27 (54) 19 (38)Epididymal sperm 10 5 (50) 2 (20)Total 85 46 (54) 21 (25)

Data are combined results from three recipient animals.*Embryos were cultured for 24 h and transferred at the two-cell stage.

Fig. 5. Lack of adenovirus integration in the mature F1 offspring. (A)Southern blot analysis of F1 DNA samples hybridized with a probe specific foradenoviral sequences (see Materials and Methods). Controls represent viralDNA in amounts equivalent to 0.1, 1, and 10 copies of viral DNA per diploidgenome. (B) PCR analysis of the F1 DNA samples. The adenovirus-specific310-bp fragment was amplified from the F1 DNA samples. Controls containing0.15 �g of normal mouse DNA were spiked with viral DNA representing 0.01,0.1, and one copy of the viral genome.

Takehashi et al. PNAS � February 20, 2007 � vol. 104 � no. 8 � 2599

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021

An important aspect of adenovirus-mediated gene delivery is itshigh transduction efficiency. Because it has been considered thatspermatogenic cells are resistant to adenovirus infection, ourcurrent results were unexpected. In the present study most of theGS cell colonies showed EGFP expression after overnight incuba-tion with AxCANEGFP; in the most successful case, 79% of thecells underwent Cre-mediated deletion. This far exceeds the effi-ciency of other methods, in which 2–30% of spermatogonial stemcells can be transduced (4, 5, 19). In contrast, previous studies havereported that adenovirus cannot infect spermatogenic cells evenwhen they are exposed to high concentrations of adenovirus (15,16). This discrepancy suggests that spermatogonial stem cells maydiffer from more differentiated spermatogenic cells in their sus-ceptibility to adenovirus infection. Given that adenovirus cannotinfect mature sperm even when they are exposed to high concen-trations of adenovirus (16), it is possible that male germ cellsacquire resistance to viral infection as they mature. On the otherhand, a potential drawback of the present approach is its toxicity.In the present study adenovirus also influenced the growth rate athigh concentrations of adenovirus, which could interfere with thegenetic manipulation of spermatogonial stem cells. Nonetheless,the infected cells were still able to differentiate normally afterspermatogonial transplantation, indicating that the virus infectiondid not compromise stem cell function.

Our results have important implications for human genetherapy. Although no evidence of stable adenovirus integrationwas found in the present study, several lines of evidence haveshown that rare integration can still occur after adenovirusinfection (at a frequency of 10�1 to 10�5 per cell) in somatic cellsand preimplantation embryos (35, 36), which raises the possi-bility that stable viral integration may also occur in germ-linecells. Given our results, it may be necessary to reevaluate thefrequency of accidental virus insertion using experimental ani-mals. In particular, the possibility of germ-line integration likelyincreases when the technique is applied to treat male infertility.We previously have rescued infertile male animals with a genetherapy approach, in which a germ cell growth factor is deliveredinto defective Sertoli cells with an adenovirus vector (13). Theinjected animals reinitiated spermatogenesis and sired offspringthat did not show viral integration. This is currently the onlymethod to rescue infertility because of Sertoli cell defects, andclinical application of this technique may rescue patients withsevere hypospermatogenesis or few germ cells. Although the lowfrequency of stable integration and preferential infection ofSertoli cells in vivo suggest that adenovirus vectors may providea relatively safe approach for treating Sertoli cell-based infer-tility, our current results indicate that caution is necessary whenextrapolating this technique to clinical cases. Further studies arenecessary to test whether such stable infection occurs in the malegerm line.

One direct application of our results in basic research is usingthis in vitro infection system for analyzing gene functions inspermatogenesis. Although several Cre transgenic lines areavailable for studying spermatogenesis, very few are available foranalysis of the spermatogonia stage, and it is often difficult tostudy gene functions in spermatogonial stem cells because theyare identified only by their function to self-renew. In this sense,adenovirus-mediated Cre expression in spermatogonial stemcells may be useful. By transplanting transgenically marked cells,it is possible to visualize the pattern and kinetics of the repopu-lation process from single stem cells (27), which can revealabnormal functions that are not evident in physiological condi-tions. The Cre infection may also be applied for selectablemarker removal after gene targeting in GS cells (37) or for moresophisticated genetic modifications. For example, a transposon/transposase construct may be used as a cargo for adenovirus, andthe ensuing transient expression of the cargo should generate anactive transposase, which then inserts the transposon cargo

permanently in the germ line. Thus, the unique mode of genedelivery by adenovirus vectors complements previously estab-lished genetic manipulation methods that achieve stable germ-line integration and will provide new opportunities for studies onmale germ cell biology.

Materials and MethodsRecombinant Adenovirus. The replication-defective adenovirusvectors AxCANLacZ and AxCANCre were obtained fromRIKEN. AxCANEGFP was a generous gift from I. Saito (Uni-versity of Tokyo, Tokyo, Japan). These vectors used the cyto-megalovirus enhancer promoter, which can be expressed inspermatogonial stem cells (32, 33). The viruses were purifiedfrom 293 cells by using CsCl centrifugation. The titer of the viruswas 2 � 108 pfu/ml, which was diluted before use.

Animals and Cell Culture. For in vivo infection, purified adenovirusvector (1 � 106 pfu/ml) was microinjected into the seminiferoustubules of ICR mice that were 5–10 days and 5 weeks old (JapanSLC, Shizuoka, Japan). For in vitro infection, adenovirus wasexposed to primary testis cells that were recovered from 6-day-old ICR mice or GS or mGS cells established from newbornDBA/2 mice (21, 22). All primary testis cells were maintained onmouse embryonic fibroblasts in Stempro34 medium (Invitrogen,Carlsbad, CA), as described (21). mGS cells were maintained inDMEM/15% FCS with leukemia inhibitory factor on gelatin-coated plates, and GS cells were cultured on laminin in Stem-pro34 medium, as described (38). In some experiments, we usedtestis cells from a R26R mouse that was kept in B6 background(The Jackson Laboratory, Bar Harbor, ME) (17). In infection ofprimary testis cells, 1 � 106 cells were plated in a 6-well plate (9.5cm2), whereas 3 � 105 cells were plated in a 12-well plate (3.8cm2) in infection of GS and mGS cells. The cells were incubatedovernight with adenovirus at concentrations ranging from 6 �103 to 6 � 105 pfu/ml. The in vitro deletion efficiency wasestimated by using homozygous R26R mice, whereas heterozy-gous R26R mice were used to examine germ-line integration inthe offspring. The cultured cells were transplanted intoWBB6F1-W/Wv (W) mice that do not have endogenous sper-matogenesis because of mutations in the c-kit gene (39).

Surgical Procedure. Microinjection into the seminiferous tubuleswas performed via the efferent duct. Each injection filled75–85% of the seminiferous tubules in each testis (40). Approx-imately 10 �l was introduced into the ICR testes, whereas only2 or 4 �l could be introduced into W mice that were 5–10 daysor 4 weeks old, respectively. When recipient mice were nothistocompatible with the transplanted cells, they were treatedwith anti-CD4 antibody to induce tolerance to the donor cells(41). The Institutional Animal Care and Use Committee ofKyoto University approved all of the animal experimentationprotocols.

Analysis of Transgene Expression. In experiments using AxCAN-LacZ, the tissues were fixed in 4% paraformaldehyde for 2 h, andX-Gal staining was used to detect LacZ expression (27). Thesame procedure was used to detect LacZ expression in R26Rmice after Cre recombination. In experiments usingAxCANEGFP, cells were analyzed by a microscope equippedwith UV fluorescence (21). For flow cytometry, the culturedcells were dissociated by trypsin, and single-cell suspensions wereincubated with rat anti-mouse EpCAM antibody (G8.8; BDBiosciences, Franklin Lakes, NJ), which was detected by allo-phycocyanin-conjugated anti-rat IgG antibody (BD Bio-sciences), as described previously (42). The cells were analyzedby a FACSCalibur system (BD Biosciences).

2600 � www.pnas.org�cgi�doi�10.1073�pnas.0609282104 Takehashi et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021

Histological Analysis. The testes were fixed with 10% neutral-buffered formalin and processed for paraffin sectioning. Twohistological sections were made from each recipient testis withan interval of 12 �m between sections. All sections were stainedwith hematoxylin and eosin.

DNA Analysis. Genomic DNA was isolated from cultured cells ortissue samples by phenol/chloroform extraction, followed by etha-nol precipitation. The deletion of the floxed allele was estimated byPCR using the 5�-TTTCTGGGAGTTCTCTGCTGC-3� and 5�-TCACGACGTTGTAAAACGACG-3� primers.

To estimate the efficiency of Cre-mediated deletion, a 270-bpfragment in the ROSA26 promoter region was amplified by PCRusing the 5�-CCTAAAGAAGAGGCTGTGCTTTGG-3� and 5�-CGTCCGGTGGAGACTTTTC-3� primers, which were used as ahybridization probe. Twenty micrograms of DNA was digested withrestriction enzymes and separated on a 1.0% agarose gel. DNAtransfer and hybridization were performed as described previously(33). To detect the adenovirus genome in offspring, a 15,043-bpScaI-EcoRI fragment of pAxCAiLacZit cosmid vector was used asa hybridization probe (Nippon Gene, Toyama, Japan). The inten-

sity of bands was quantified by NIH Image 1.63. To detect virusintegration, a 310-bp region of adenovirus type 5 was amplified byPCR using specific primers, as described previously (13).

Microinsemination. The seminiferous tubules of recipient micewere mechanically dissected, and spermatogenic cells werecollected. Microinsemination was performed by intracytoplas-mic injection, as described previously (28). Embryos thatreached the two-cell stage after 24 h in culture were transferredto the oviducts of day-1 pseudopregnant ICR female mice.Fetuses that were retrieved on day 19.5 were raised by ICR fostermothers.

We are grateful to Ms. A. Wada for her technical assistance. We alsothank Dr. M. Ikegawa for providing us with pAxCAiLacZit cosmidvector. Financial support for this research was provided by the Ministryof Education, Culture, Sports, Science, and Technology of Japan and bygrants from CREST and the Human Science Foundation (Japan). Thiswork was also supported in part by the Tokyo Biochemical ResearchFoundation, the Genome Network Project, and Special CoordinationFunds for Promoting Science and Technology from the Ministry ofEducation, Culture, Sports, Science, and Technology.

1. de Rooij DG, Russell LD (2000) J Androl 21:776–798.2. Meistrich ML, van Beek MEAB (1993) in Cell and Molecular Biology of the

Testis, eds Desjardins C, Ewing LL (Oxford Univ Press, New York), pp266–295.

3. Nagano M, Shinohara T, Avarbock MR, Brinster RL (2000) FEBS Lett475:7–10.

4. Nagano M, Brinster CJ, Orwig KE, Ryu B-Y, Avarbock MR, Brinster RL(2001) Proc Natl Acad Sci USA 98:13090–13095.

5. Hamra FK, Gatlin J, Chapman KM, Grellhesl DM, Garcia JV, Hammer RE,Garbers DL (2002) Proc Natl Acad Sci USA 99:14931–14936.

6. Wivel NA, Gao G-P, Wilson JM (1999) in The Development of Human GeneTherapy, ed Friedmann T (Cold Spring Harbor Lab Press, Plainview, NY), pp87–110.

7. Shui J-W, Tan T-H (2004) Genesis 39:217–223.8. Neering SJ, Hardy SF, Minamoto D, Spratt SK, Jordan CT (1996) Blood

88:1147–1155.9. Blanchard KT, Boekelheide K (1997) Biol Reprod 56:495–500.

10. Ye X, Gao GP, Pabin C, Raper SE, Wilson JM (1998) Hum Gene Ther9:2135–2142.

11. Paielli DL, Wing MS, Roguiski KR, Gilbert JD, Kolozsvary A, Kim JH, HughsJ, Schnell M, Thompson T, Freytag SO (2000) Mol Ther 1:263–274.

12. Boyce N (2001) Nature 414:677.13. Kanatsu-Shinohara M, Ogura A, Ikegawa M, Inoue K, Ogonuki N, Tashiro K,

Toyokuni S, Honjo T, Shinohara T (2002) Proc Natl Acad Sci USA 99:1383–1388.

14. Kojima Y, Sasaki S, Umemoto Y, Hashimoto Y, Hayashi Y, Kohri K (2003)J Urol 170:2109–2114.

15. Peters AHFM, Drum J, Ferrell C, Roth DA, Roth DM, McCaman M, NovakPL, Friedman J, Engler R, Braun RE (2001) Mol Ther 4:603–613.

16. Hall SJ, Bar-Chama N, Ta S, Gordon JW (2000) Hum Gene Ther 11:1705–1712.17. Soriano P (1999) Nat Genet 21:70–71.18. Shinohara T, Orwig KE, Avarbock MR, Brinster RL (2001) Proc Natl Acad Sci

USA 98:6186–6191.19. Kanatsu-Shinohara M, Toyokuni S, Shinohara T (2004) Biol Reprod 71:1202–

1207.20. Anderson R, Schaible K, Heasman J, Wylie CC (1999) J Reprod Fertil

116:379–384.

21. Kanatsu-Shinohara M, Ogonuki N, Inoue K, Miki H, Ogura A, Toyokuni S,Shinohara T (2003) Biol Reprod 69:612–616.

22. Kanatsu-Shinohara M, Inoue K, Lee J, Yoshimoto M, Ogonuki N, Miki H,Baba S, Kato T, Kazuki Y, Toyokuni S, et al. (2004) Cell 119:1001–1012.

23. Brinster RL, Zimmermann JW (1994) Proc Natl Acad Sci USA 91:11298–11302.

24. Kanatsu-Shinohara M, Inoue K, Miki H, Ogonuki N, Takehashi M, MorimotoT, Ogura A, Shinohara T (2006) Biol Reprod 75:68–74.

25. Nagy A (2000) Genesis 26:99–109.26. Kanegae Y, Lee G, Sato Y, Tanaka M, Nakai M, Sakaki T, Sugano S, Saito I

(1995) Nucleic Acids Res 23:3816–3821.27. Nagano M, Avarbock MR, Brinster RL (1999) Biol Reprod 60:1429–1436.28. Kimura Y, Yanagimachi R (1995) Development (Cambridge, UK) 121:2397–

2405.29. Palermo G, Joris H, Devroey P, Van Steirteghem AC (1992) Lancet 340:17–18.30. Schaller J, Glander HJ, Dethloff J (1993) Hum Reprod 8:1873–1878.31. Wickham TJ, Mathias P, Cheresh DA, Nemerow GR (1993) Cell 73:309–319.32. Niwa H, Yamamura K, Miyazaki J (1991) Gene 108:193–199.33. Kanatsu-Shinohara M, Toyokuni S, Shinohara T (2005) Biol Reprod 72:236–

240.34. Nagano M, Watson DJ, Ryu B-Y, Wolfe JH, Brinster RL (2002) FEBS Lett

524:111–115.35. Harui A, Suzuki S, Kochanek S, Mitani K (1999) J Virol 73:6141–6146.36. Tsukui T, Kanegae Y, Saito I, Toyoda Y (1996) Nat Biotechnol 14:982–985.37. Kanatsu-Shinohara M, Ikawa M, Takehashi M, Ogonuki N, Miki H, Inoue K,

Kazuki Y, Lee J, Toyokuni S, Oshimura M, et al. (2006) Proc Natl Acad Sci USA103:8018–8023.

38. Kanatsu-Shinohara M, Miki H, Inoue K, Ogonuki N, Toyokuni S, Ogura A,Shinohara T (2005) Biol Reprod 72:985–991.

39. Geissler EN, Ryan MA, Housman DE (1988) Cell 55:185–192.40. Ogawa T, Arechaga JM, Avarbock MR, Brinster RL (1997) Int J Dev Biol

41:111–122.41. Kanatsu-Shinohara M, Ogonuki N, Inoue K, Ogura A, Toyokuni S, Honjo T,

Shinohara T (2003) Biol Reprod 68:167–173.42. Shinohara T, Avarbock MR, Brinster RL (1999) Proc Natl Acad Sci USA

96:5504–5509.

Takehashi et al. PNAS � February 20, 2007 � vol. 104 � no. 8 � 2601

APP

LIED

BIO

LOG

ICA

LSC

IEN

CES

Dow

nloa

ded

by g

uest

on

Aug

ust 2

0, 2

021