charles thibault - edp sciences

Charles Thibault

Throughout his life, Charles Thibault wasalways fascinated by research. During the1950’s, at the Institut National de la Recher-che Agronomique, he developed one of thefirst laboratories of the Animal PhysiologyDepartment, of which he later became theDepartment head. Later, he had other re-search administrator tasks as the Presidentof the Centre National de la RechercheScientifique. As a Professor of ReproductionPhysiology at the Université Paris 6, he wasalways attentive to include the most recentresearch results in his classes. He was advisorfor a very large number of doctorate studentsand many students owe him for helping themto become researchers. But above all, he wasfascinated by research and was creative. Hewas responsible for the first in vitro fertilisa-tions in domestic mammals. His scientificknowledge and his intellectual vivacity al-ways impressed his colleagues in the manycongresses in which he participated and hewas, very justly, honoured with many dis-tinctions. But he also knew how to walk outof his laboratory and take part in debates ofthe society against, notably, narrow mindedpositions on birth regulation and systematicmedicalization of human procreation. Fasci-nated by scientific writing and publication,the Professor Charles Thibault spent a large

part of his activities publishing and diffusinghis research results. In 1961, with AndréFrançois he founded the journal “Annales deBiologie Animale, Biochimie, Biophysique(ABABB)”, mainly dedicated to papers inanimal physiology and nutrition biochemis-try submitted by researchers at INRA. Thisjournal rapidly became a success and itspages were quickly opened to other researchorganisations. Charles Thibault, when hand-ing over his position as Editor-in-Chief of thejournal in 1981, recognised the importanceof reorientating the journal, which he madepublic by a change in its title (Reproduction,Nutrition, Development), and increasing itsreadability and diffusion. He was an intransi-gent negotiator with the Editors, a respectedcounsellor to the authors, and he was alwayspreoccupied by the quality of the journal, notonly with its scientific selection of the manu-scripts but also with their presentation. He al-ways found an undissimulated pleasure inletting his fingers walk through the latest is-sue of the journal and “touching the paper”that was a product of several months ofshared work. In publication as elsewhere, hiscandour, his enthusiasm and his scientificrigour led him to be wary of preconceivedideas and to renounce diverse theories thatwere not supported by incontestable experi-mental results. His colleagues still rememberthe energy he spent directing the new editionof “Reproduction chez les Mammifères etl’Homme” in 2001, the only French work inthis domain that has been published recently.Our journal is honoured to publish here hislast review article on human cloning.

The editors of RND regret that this mas-ter who they profoundly respected and whowas always natural and spontaneous, will nolonger be around to give them council andencouragement.

The old and current Editorsof Reproduction Nutrition Development

Review

Recent data on the development of cloned embryosderived from reconstructed eggs with adult cells

Charles THIBAULT*

UMR Biologie du Développement et Reproduction,Institut National de la Recherche Agronomique, 78352 Jouy-en-Josas Cedex, France

(Received 3 February 2003; accepted 15 June 2003)

Abstract — Production of cloned embryos by nuclear transfer from adult somatic cells is a noveland promising technique in animal biotechnology. In spite of numerous reported viable offspring invarious species, the efficiency of the technique remains very low. Embryonic and fetal mortality oc-curs all along pregnancy and during the peri-natal life, even months after birth. Both embryonic andplacental dysfunctions might be involved. However the precise causes of such developmental fail-ures are still unknown. In the present review, we report data from different studies which describedthe main defaults which have been observed after embryonic cloning in various species. The puta-tive molecular and cellular causes of these developmental failures are discussed.

cloning / development failure / nuclear reprogramming

1. GENERAL SURVEY

Many mammals have been cloned butthe ratio between the number of recon-structed embryos transferred and the num-ber of viable offspring, remains very low, inspite of a broad spectrum of investigations.Most of the embryos die during the preg-nancies, and even after parturition. It israther puzzling, however, those that surviveare physiologically normal as shown [1] in alarge study on bovine cloning. From 2170recipients receiving 1 to 3 cloned embryos,117 calves were born (5.4%), 82 survivedand developed as adults (3.8% of all preg-nant recipients, 2% of the total number of

embryos transferred). Cumulative abortionrates reach 75%. Interestingly, prior toabortion, the growth rate of aborted fetuseswas similar to those that were not aborted.This observation suggests that the trophecto-derm and then the placenta are mainly re-sponsible for embryonic death. However,since many offspring die after parturition,embryo deficiencies are also a cause ofabortion.

Surviving cloned heifers started to displaysigns of reaching puberty at 10–11 months.All of the 22 cloned heifers that were insem-inated became pregnant. Age at calving was23–25 months, similar to non-cloned cattle.The milk from cloned cattle appeared to be

* Deceased on 20th August 2003.Corresponding author: Michel Guillomot: [email protected]

Reprod. Nutr. Dev. 43 (2003) 303–324 303© INRA, EDP Sciences, 2003DOI: 10.1051/rnd:2003027

normal. An aberrant epigenetic modifica-tion (large offspring syndrome) has, how-ever, been observed in living cloned mice,although fertile [2].

There is a general agreement first thatembryonic mortality increases with the ageof the donor from the embryo to the adult,and second that embryonic mortality per-sists all along gestation. An example isgiven in Table I.

This table clearly shows that the late fetallosses (in ital.) from day 90 to day 282 arehigh when donor cells are from adult ani-mals [3].

Renard et al. [4] have given a very com-plete overview of the biological aspects ofcloning and clones. They give new data,mainly devoted to cattle, and discuss on thepresent results and future prospects.

The present review focuses attention onthe causes of abnormal development ofcloned embryos.

2. NUCLEAR REPROGRAMMING

2.1. Role of GV content

At what stage of maturity are oocytes ableto reprogram a somatic nucleus? Gao et al.[5] studied the capacity of immature (GV)

or premature (pro-M I) mouse cytoplasm toreprogram a nucleus of an embryonic stemcell. Control oocytes were in M II beforenuclear transfer (Tab. II).

Thirty minutes after the nuclear transferin all three types of cytoplasm, a spindlestarts to organize and condensed chromo-somes are formed. In GV ooplasm, chromo-somes appear distributed into the cytoplasmand are neither organized on the spindle norform a pronucleus. In contrast, in Pro-M Icytoplasm (GVBD stage), condensed chro-mosomes reach the spindle and form ametaphase plate similar to the control M IIas shown by the immunofluorescence tech-nique, and after activation a pronucleus isformed. Only 8.1% of pro-M I reconstructedembryos developed to the morula-blastocyststage. In contrast, 53.5% of M II recon-structed embryos reached this stage and6 pups are recovered from the surrogatemothers at 19.5 days post coitum. Moreover,during the 5 h following transfer and activa-tion, the content of microfilaments in the cor-tex of GV oocytes (as well as Pro-M Ioocytes) was lower than in the cortex of M IIooplasm. It seems that GV material is essen-tial for nuclear remodelling after nucleartransfer: In GV-ooplasm, after activation, re-constructed oocytes could not form pseudo-pronuclei, and the chromosomes remained

304 C. Thibault

Table I. Evolution of 240 cattle pregnancies after transfer of day 7 blastocysts according to the originof the cell donor (from Heyman et al. [3]).

Pregnant recipientsat the followingdays or months

Adult somaticcell

133 recipients

Fetal somaticcell

40 recipients

Embryoniccell

67 recipients

In vitro fertilizedembryos

51 recipients

Day 21Day 35Day 50Day 70Day 903–5 months5–7 months7–9 monthsNo. calves born (%)

74/13345/13336/13319/13316/13313/13310/1339/133

9/133 (6.8%)

23/4011/409/409/409/409/407/406/40

6/40 (15%)

42/6733/6728/6725/6723/6723/6722/6722/6723/67

(34.3%)

32/5127/5126/5125/5124/5124/5124/5124/51

25/51 (49%)twin calving

condensed. The factor(s) released by GV, af-ter GVBD, are unknown but seem in relationwith chromosomes and spindle recognitionand with nucleus reconstruction.

2.2. Chromatin remodeling

Three types of recipient enucleatedoocytes have been used for nuclear transfer:non-activated oocytes (TA), previously ac-tivated oocytes (AT) and ageing oocytes inwhich the MPF level is decreasing.

Taking normal development as a proofof nuclear reprogramming, Tani et al. [6]examined the in vitro developmental poten-tial of TA and AT enucleated bovine ova re-ceiving cumulus cell nuclei at various stagesof the cell cycle: cycling, G0/G1, G1, S, G2and M.

The development potential of nucleartransfer ova was clearly different dependingon the condition of the recipient ova.

When activated ova were used as therecipient cytoplasm, none of the nuclear-transferred ova receiving cumulus cell nu-clei at any cell cycle stage, developed be-yond the 8-cell morula stage. In contrast, 21to 50% of non-activated ova receiving cu-mulus cell nuclei at all cell-cycle stages,except the S phase, and then activated, de-veloped into blastocysts.

The blastocyst cell number was not sig-nificantly different among any group.(However, the numbers of cells given inTab. I for 8-day blastocysts (from 70.7 ±29.4 to 86.3 ± 31.8) are much lower thanthose found in blastocysts obtained fromin vitro or in vivo fertilization (≥ 200 cells).)

Their studies demonstrate that aftertransfer of a somatic nucleus in a non acti-vated ooplasm, the envelope of the somaticnucleus is broken and chromosomes are di-rectly exposed to ooplasm allowing nuclearremodeling. In contrast, in activated ooplasmthe nucleus remains unchanged.

Kim et al. [7] has made a similar conclu-sion in a mouse nuclear transfer experiment:there is 1% of morulae/blastocysts when cu-mulus cell nuclei are transferred after acti-vation (AT) vs. 38.3% when transfer occursbefore activation (TA). Moreover, theycompared transcriptional activity and chro-mosome DNAse-1 sensitivity in TA and ATreconstructed ova. A basal transcription fac-tor (TATA box binding protein = TBP) wasdisplaced from the condensed chromatin oftransferred nuclei in the TA ooplasm. Suchtranslocation of TBP was not observed inthe AT embryos. In the TA embryos,DNAse-1 sensitivity was increased after nu-clear transfer (2-fold, at 5 h postactivation),then decreased at 12 h, while the sensitivitywas not changed in AT embryos during thefirst-cell stage. The results suggest that the

Development of cloned embryos 305

Table II. Ooplasm status and per cent of blastocysts and fetuses obtained after nuclear transfer (Gaoet al. [5]).

Enucleationstage

No. oocytessurviving

injection andactivation

No. and(%)

activatedeggs

No. and(%)

cleavedeggs

No. and(%) morula /blastocystsdeveloped

No.transferredembryos

(recipients)

GV

Pro-M I

M II

75

86

221

0

85 (98.8)

213 (96.4)

42 (48.8)

162 (76.1)

7 (8.1)

114 (53.5)* 114 (11)+

+ 6 pups were recovered from surrogate mothers at 19.5 days postcoitum.*P < 0.01.

cytoplasm of activated mouse oocytes isinefficient for nuclear remodeling due toa failure of nuclear envelope breakdownwhich allows direct interaction of chromo-somes with cytoplasmic factors.

2.3. Structure of the DNAof the transferred nucleus

Loi et al. [8] studied the resistance ofDNA and centrioles of ovine granulosa cells(GC) heated to lethal temperatures. All nu-clear proteins were denatured at 55 or 75 oCand GC became nonviable, however, DNAremained able to be reprogrammed (Tab. III),(a functional centriole was postulated butnot shown).

Unfolded and denatured proteins are morereadily degraded by the oocyte proteolytic en-zymes. Therefore, the oocyte reprogramming

machinery might more readily processnucleoprotein complexes rendering the chro-mosomes more accessible to remodeling.However, the results remain low althoughslightly better than those obtained in ewes af-ter nuclei somatic cell transfer (Tab. IV).

2.4. Imprinting/epigeneticreprogramming. Methylation

2.4.1. Embryos from fertilized eggs

Both the sperm and egg genomes arehighly methylated and transcriptionally si-lent at the time of fertilization. Within only afew hours after fertilization, the paternal ge-nome of the mouse embryo undergoes adrastic and active demethylation, whereassimilar genome-wide demethylation of thematernal genome occurs progressively after

306 C. Thibault

Table III. Heat denaturation of granulosa cell nucleoproteins and subsequent development after nu-clear transfer (Loi et al. [8]).

Nuclei from Injected (n) Cultured (n) Blastocysts (n and %)

GC, untreated

GC, heated at 55 °C

GC, heated at 75 °C

Fertilized

241

470

242

–

227

336

146

335

25 (11%)

47 (14%)

41 (28%)

136 (40.5%)

Table IV. The percentage of pregnant ewes decreased similarly whatever the treatment. However, nolamb was obtained with granulosa cells (GC) heated to 75 °C (Loi et al. [8]).

Nuclei fromBlastocysts transferred

number of recipientewes (number of ewes)

No. and (%) of ewespregnant Lambs

(n and %)at day 40 at day 60 at day 80

GC untreated

GC, 55 °C

GC, 75 °C

Fertilized

25 (13)

40 (19)

25 (11)

52 (26)

9 (69%)

15 (79%)

7 (64%)

20 (77%)

5 (38%)

9 (47%)

5 (45%)

18 (69%)

2 (15%)

5 (26%)

3 (27%)

18 (69%)

2 / 25 (8%)

4 / 40 (10%)

0

23/52 (44%)

the two-cell embryonic stage in the mouse.Both paternal and maternal genomes areequivalently under-demethylated at themorula and blastocyst stages and, thereafter,de novo methylated [9].

From the beginning of embryonic devel-opment, some genes, called imprintedgenes, are inactivated either in the male or inthe female genome by high methylation.

Kierszenbaum [10] has summarized ourknowledge on imprinting and the occur-rence of reprogrammation of imprintedgenes after somatic nuclear transfer.

Genomic imprinting, an epigenetic formof gene regulation, determines the par-ent-dependant gene expression of marked orimprinted genes during gametogenesis andembryonic development. About 50 mam-malian imprinted genes have been identi-fied. Imprinting involves differential alleleDNA hypermethylation in one sex cell lin-eage but not in the other. Primordial germcells inherit biallelically imprinted genesfrom paternal and maternal origin and erasethese imprints to start de novo monoallelicimprinting during gametogenesis.

Epigenetic paternalization is an ongoingprocess in the mitotically-dividing sperma-togonial stem cell and is derived meioti-cally, dividing spermatocyte progeny toendow spermatozoa with imprinted alleles.Epigenetic maternalization is restricted tothe oocyte growth phase and is unrelated toDNA replication since it takes place whilethe oocyte remains in the diplotene stage ofmeiotic prophase I.

Genomic methylation depends on the ac-tivity of DNA methyl transferases (DnmT).Methylation takes place mainly on the cystineresidues that are present in dinucleotide mo-tifs consisting of a 5’ cytosine followed by aguanosine (CpG). DNA methylation pat-terns are regulated during development bythree distinct methyltransferases, Dnmt-1,Dnmt-3a and Dnmt-3b. Dnmt-1 is essen-tially a house-keeping methyltransferaserequired for maintaining tissue specific

methylation patterns as shown by mutationsin the Dnmt-1 gene, which result in globalgenome demethylation and embryo lethality.Dnmt-3a and Dnmt-3b are required for denovo methylation of embryonic stem cellsand early post implantation embryos.

How do these methyltransferases recog-nise the target sequences in candidate genesin the male and female genome?

A partial answer has been given by thediscovery of the role of a protein sharinghomology with methyltransferases but lack-ing enzyme activity, the Dnmt 3L. Hata et al.[11] have shown that this protein seems nec-essary for targeting methyltransferases toDNA specific sequences. Dnmt3L expres-sion was detected in oocytes, in ES cells andin the chorion. The expression pattern ofDnmt3L in all these cells and tissues is strik-ingly similar to that of Dnmt3a and Dnmt3b.Dnmt-3L–/– female mice had normal repro-ductive functions but no live pups wereborn. They died around E10,5 with head andneural abnormalities and under developedplacentas. Although Dnmt3L had no methyl-transferase activity, the methylation statusof differentially methylated regions of somematernally imprinted genes were almostcompletely demethylated. The same geneswere also unmethylated in Dnmt3a–/– andDnmt3b–/– mice.

Unlike the female Dnmt3L–/– mice, themethylation of male imprinted genes, H19and Rasgrf1, were unaffected. However, theDnmt3L–/– male was sterile: the spermato-gonia failed to differentiate into spermato-cytes: the lack of Dnmt3L appears toprevent spermatogonial progeny from un-dergoing differentiation and meiosis.

2.4.2. Cloned embryos

To get a better understanding of globalmethylation differences associated with thedevelopment of nuclear transfer generatedcattle, Cezar et al. [12] analyzed the ge-nome-wide methylation status of 9 spon-taneous aborted cloned fetuses (2-, 2.5-, 6-,


7-month-old), 7 cloned normal fetuses re-covered from ongoing normal pregnancies(# 2 months old), 4 IVF foetuses (1.5-,2-months old), 6 fetuses generated by artifi-cial insemination (2-months old), 11 adultclones and 4 adults from natural reproduc-tion (Tab. V).

The results indicated an extensive hypo-methylation of DNA in the spontaneouslyaborted clones (P < 0.001) and a significant(P < 0.005) reduced methylation in livingcloned fetuses compared with in vivo fe-tuses. There was no difference betweenadult clones and in vivo adults (Tab. IV).This implies that the survivability of clonedcattle may be closely related to the globalmethylation status.

Excessive growth in cloned living ani-mals might be due to a wrong imprinting ofgenes involved in body weight regulation.Tamashiro et al. [2] studied the fate of clonedmice which remained alive, but showedgrowth abnormalities (large offspring syn-drome and obesity). Leptin and insulin levelswere much higher than in the control (7- and3-fold respectively) but similar to those ofspontaneous obese mice in which either theob gene or its db receptor was mutated. How-ever, the cloned mice were fertile and theiroffspring were of normal body weight. Thisindicates that gene dysfunction (hypome-thylation of one parental allele) in clonedmice might be of epigenetic nature.

3. CAUSES OF EMBRYONICAND FETAL DEATHS:MORPHOLOGICALABNORMALITIES,AND CELLULAR AND GENETICDEFECTS

Extensive studies have been developedto find out the causes of embryonic loss andto overcome the discouraging percentage ofliving young obtained, whatever the cloningprocedures used. At present, it seems thatdefective mechanisms take place very earlyin development, although their conse-quences may occur along pregnancy andeven some days or weeks after birth. Com-parative studies between embryos from nu-clear transfer (NT), in vitro fertilization(IVF), intra cytoplasmic sperm insemina-tion (ICSI), parthenogenesis, and in vivofertilized and developed embryos (IFD =in vivo fertilization and development), wereperformed to possibly identify the differ-ences during the beginning of development.

3.1. Growth of the inner cell massand trophectoderm

Rybouchkin et al. [13] studied the poten-tial of the development of cloned mouseembryos in comparison with the potential ofICSI embryos and parthenogenetic embryos.They demonstrated that the capacity of re-constituted eggs to develop to the blastocyst

308 C. Thibault

Table V. Genomic cytosine methylation content of fetal and adult genomes detected by reverse phaseHPLC (Cezar et al. [12]).

Group Number of animals% of

5mC /(5mC + C)*

AbortedCloned fetusesIn vivo fetusesIVF fetusesAdult clonesIn vivo adult

9764

114

P < 0.0017

2329273229

P < 0.005

* Results were expressed in % of methyl cytidine/total cytidine.

}

{

stage (29%) was significantly lower thanthat of ICSI or parthenogenetic embryos(95% and 92% respectively). Clonedblastocysts had a significantly lower meannumber of cells in the ICM (9) andtrophectoderm (52) than did their counter-parts (31 and 143 for ICSI; 21 and 92 forparthenogenetic blastocysts). This corre-lated with a significantly higher proportionof dead cells in the cloned blastocysts.

Koo et al. [14] studied allocations of thetwo different cell lineages of the bovineblastocysts, the inner cell mass (ICM), andthe trophectoderm (TE).

Total cell numbers as well as the numberof TE cells, were significantly smaller inIVF and NT blastocysts than in control invivo blastocysts of the same age while ICMcell numbers were identical (Tab. VI).These results suggest that NT embryos hav-ing fewer TE cells at the blastocyst stagemight form smaller placentas, eventuallyleading to fetal loss (Tab. VI).

3.2. Abnormal developmentof the placenta

A high incidence of placental abnormali-ties was observed during gestation: placen-tas were either overgrown (mouse), under-developed (bovine, sheep) or abnormallydeveloped.

3.2.1. Overgrowth

In the mouse, Wakayama and Yanagimachi[15] compared the weight of placentas atterm from “cloned” fetuses and “in vivo fer-tilized” fetuses. There were differences ac-cording to the origin of the donor cells(cumulus, ovary or testis, skin, and ES cells)between 0.34 ± 0.07 g to 0.19 ± 0.07 g butwhatever the origin of the donor cell, theseweights were significantly higher (P < 0.01)than the placenta of control “in vivo fertil-ized” fetuses (0.15 ± 0.03 g for male and0.12 ± 0.02 g for female fetuses). Ono et al.[16] gave similar values (0.36–0.29 g and0.13 g). This placentomegaly is caused byan expansion of the spongiotrophoblast withan increased number of glycogen cells andan enlargement of trophoblastic cells. Onoet al. [16], Ogura et al. [17], Suemizu et al.[18], and Tanaka et al. [19] observed pla-cental abnormalities from embryos that werederived through other micromanipulationtechniques (ICSI, aggregation chimeras,pronuclear exchange) but expansion of thebasal layer with marked proliferation of gly-cogen cells is the only phenotype that isunique to cloning. Expression of 3 imprintedgenes (IGF-2 (paternally expressed), IGF-2rand P57Kpi2 (maternally expressed)) is slightlyreduced apparently due to the reduced pro-portion of the labyrinth layer and not to achange in transcriptional activity [19]. Why


Table VI. No. of cells in 7-day-old bovine blastocysts from in vivo and in vitro fertilization, andnuclear transfer (Koo DB et al. [14]).

GroupNo. of

blastocystsexamined

No. of cells

ICM TE Total ICM / ICM + TE

In vivo

In vitro

NT

26

52

55

42.3 ± 10.5

45.6 ± 18.4

48.6 ± 25

80.5 ± 19.5

62.4 ± 26.3**

49.5 ± 1.5*

122.5 ± 21.6

108.2 ± 33.3**

98.0 ± 44.3**

34.9 ± 8.9

42.6 ± 14.9**

50.1 ± 17.9**

* P < 0.01; ** P < 0.05.

and how the spongiotrophoblast layer ex-pands, remains unclear.

In cloned mice, most concepti were lostby 10 days of pregnancy, that is before thehemochorial placenta starts to function inthe mouse. Given this fact, the late placentaphenotype that these authors describedseems unlikely to be a cause of the embry-onic mortality.

Inoue et al. [20] have shown that whilethe levels for imprinted genes Igf2, Igf2rand H19 were within the control range,there was a significant reduction of mRNAlevels of three imprinted genes and fournonimprinted genes in the placenta of bothSertoli cell- and cumulus cell-derived clonescompared with the controls. In contrast, thecorresponding cloned fetuses exhibitedsteady state mRNA levels for these geneswithin the control range. Thus the placentais more affected than the fetus.

3.2.2. Hypotrophy

In cattle, at day 30 of pregnancy,histological examination reveals the pres-ence of placentomes and a tenuous attach-ment of maternal and fetal epithelia. At thisperiod, Hill et al. [21] found no significantdifference between pregnancy rates for re-cipients that carried cloned fetuses and re-cipients carrying in vivo or in vitro fertilizedfetuses. In cloned pregnancies, a high levelof embryonic deaths occurs after day 30.Whereas 100% of the control fetuses sur-vived at day 90 only 19% of the clonedfetuses present on day 30, survived. Out of6 placentas from NT embryos, 4 wereabnormally developed (hypotrophy, poorvascularization, rudimentary cotyledons).

3.2.3. Abnormal placentation

In cattle, placentas from clones undergo-ing an abnormal pregnancy show a lowernumber of placentomes than those from invitro fertilization (69 versus 99) but theirmean weight is higher: 144.3 ± 21.4 g in

abnormal cloned fetuses and 72.9 ± 4.8 g innormally developed cloned fetuses versus33.6 g in two normal in vitro produced fe-tuses [22]. The large placentomes areoedematous. Studies by ultrasonographyduring gestation allow to visualize theselarge placentomes and to predict abortion [3].

Reduction of the numbers of placen-tomes seems to indicate that the placenta atthe time of implantation possesses an insuf-ficient number of molecular sites to recog-nize all uterine caruncles.

3.3. Immunological status

Hill et al. [21] studied the potential foraltered immunological status of clonedpregnancies to be a contributing factor ofembryonic loss. Expression of the majorhistocompatibility complex class I (MHC-I)by trophoblast cells and distribution of endo-metrial T-lymphocytes were investigated.

In cattle, at the beginning of a naturalpregnancy when the intimate contact be-tween endometrium and chorioallantois isinitiated, MHC-I expression is normallysuppressed. In contrast, in eight clonedpregnancies, all derived from the same fetalcell line, and examined at 5 weeks (6),7 weeks (1) and 8 weeks (1), all 8 clonedplacentas displayed trophoblast MHC-I ex-pression which varied within clones, from17.9 to 56.5% of the trophoblast cells. Noneof the 8 controls (4–7 weeks old) showedMHC-I expression.

In the cloned pregnancies the numbers ofT-lymphocyte CD3+ were higher in theendometrium compared with the controls.Moreover, large aggregates of T-cells werefrequently observed in addition to increasednumbers of diffusely spread sub-epitheliallymphocytes. The observed expression ofMHC I in cloned pregnancies is likely tohave induced a maternal lymphocytic re-sponse that would be detrimental to main-taining the cloned pregnancies: 4 alive,3 dead at days 34–49 of pregnancy versus

310 C. Thibault

7 alive, 1 dead for the controls at days 29–45(estimation).

An abnormal immunological status mayalso be observed in a young cloned calf afterbirth, with a sudden and rapid fall in lym-phocyte count due to thymic atrophy 6–7weeks after birth, leading to death on day 51(Renard et al. [23]). Mice cloned from im-mature Sertoli cells, and bred in rigorouslycontrolled pathogen-free-conditions diedprematurely from pneumonia and reducedproduction of antibodies compared with thatof the controls [17]. Immuno-deficiency hasbeen mentioned incidentally in some otherexperiments.

3.4. Telomere length

There is a general agreement thattelomere length regresses with age. Ques-tions have been raised whether organismscreated by nuclear transfer with cells from

adult and aged animals, will undergo pre-mature senescence. In sheep, Shields et al.[24] observed that in three one-year oldsheep cloned with mammary cells from a6-year old ewe, the telomeres remainedshorter compared with the aged-matchedanimals (terminal restriction fragments(TRF) from 19.14 kb to 20.37 kb versus23.9 ± 0.18 kb. In contrast to sheep, cloningresets the “telomere clock”1 in cattle andmice. Lanza et al. [25] have successfullycloned with senescent bovine fibroblasts.Fibroblasts from embryonic tissue, werepassaged until greater than 95% of theirlife-span were completed. They obtainedthe birth of 6 calves (out of 1896 nucleartransfers) (Tab. VII).

At senescence, their TRF was decreasedfrom 18.3 kb at the beginning of the culture,


1 “Telomere clock” during the life cycle by analogywith the circadian clock.

Table VII. Changes in cell telomere length of transferred nuclei, in somatic cells nuclei of cloned cat-tle (Lanza RP et al. [25]).

Donor cell type

Mean telomererestriction fragment(kb) in transferred

nucleus

Mean telomererestriction fragment

in cloned cattlecells (kb)

Telomere lengthchange (kb)

Oviductal epithelialcells from a 13-yr-oldHolstein cow

Mammary epithelialcells from a 13-yr-oldHolstein cow

Oviductal epithelialcells from a 6-yr-oldJersey cow

Muscle cells from a12-yr-old JapaneseBlack bull

Skin fibroblasts froma 2-yr-old JapaneseBlack bull

15.2

16.3

16.9

20.1

18.2

13 (12.5 to 13.7)

14.3

15.5 (14.9 to 16.0)

19.7 (19.6 to 19.9)

20.2 (20.0 to 20.4)

–2.1(–1.5 to –2.7)

–2.0

–1.5

–0.3

+2.0

to 15.2 kb at the end. In 7 week-old embryoscloned with these senescent fibroblasts,telomere length was increased to 20.1 kband the replicative life-span of cells fromthese embryos was 90 ± 1.6 populationdoublings versus 60.5 ± 1.7 in the primaryculture. Finally, the telomeres of 5–10 monthold cloned cattle were longer than thetelomeres of age-matched control animals.Thus in cattle, cloning over resets telomerelength and restores a youthful state. Similarresults have been published by Tian et al.[26] who used cells from a 13-year-old cow.All 4 living calves had longer telomeresthan the donor cow (15.38 ± 0.2 kb versus12.43 ± 0.49 kb; P < 0.05), and the telomerelength of cloned calves was not differentfrom the telomere length of their age-matched controls (15.38 ± 0.2 kb versus14.73 ± 0.49 kb). Miyashita et al. [27] fo-cused their attention on the differences intelomere lengths according to the tissues;they produced reconstructed embryos withnuclei from oviduct or mammary gland epi-thelial cells, from skin fibroblasts or musclecells.

Their results are summarized in Ta-ble VII. In control Japanese Black males,TRF regresses from 20.5 kb at birth to 16 kbin two 18 year-old bulls of the same breed.

It seems that telomere length decreaseswhen donor cells are oviduct epithelialcells, mammary gland and muscle cells. Incontrast, telomere length increases whendonor cells are ear-skin fibroblasts. Atten-tion might be drawn on the similarity oftelomere reduction in calves cloned withfibroblasts from the cow mammary gland,as shown in sheep.

However, the choice of three differentbreeds of cattle weakens these conclusions.

3.5. Aberrant patternsof X chromosome inactivation

In mice and in cattle, X-inactivation isimprinted in the placenta, in which the pa-ternal X is preferentially inactivated.

Xue et al. [28] studied allele specificexpression of the X-linked monoamineoxydase and nine additional X-linked genesin 9 cloned XX calves. Six calves were bornfrom corona cell nuclear transfer, 4 survivedand remained healthy whereas the other twodied. In contrast to control calves, theyfound random X-inactivation in the placentaof the deceased cloned embryos and a com-pletely skewed X-inactivation in that of liveclones. They found aberrant expression pat-terns in nine of the 10 linked genes. In oneclone, Xist expression was missing.

Since hypermethylation and hypomethyl-ation of Xist were correlated with the X inac-tivated and X-activated chromosomesrespectively, they examined the methylationpatterns of the 5’ region of bovine Xist bymethylation sensitive PCR. Different de-grees of hypomethylation were detected inthe hearts in two clones, and in the liver andspleen of one clone.

They believe that the aberrant X-inacti-vation patterns might result from the factthat somatic cloning bypasses the completereprogramming process which normally oc-curs in preimplantation embryos and ingerm cells. Consequently epigenetic signalssuch as DNA methylation, histone acetyl-ation and chromatin structure may or maynot be correctly reprogrammed during theprocess of nuclear remodeling in cattle. In-formation was missing on the state ofmethylation of normally imprinted genesimmediately after nuclear transfer and dur-ing the beginning of cleavage (see Reik [9]).

3.6. Donor cell types

The efficiency of cells from many tissues(cumulus, oviduct, uterus, liver, skin, fromadult, newborn or embryos) has been stud-ied mainly in cattle for two reasons: the eco-nomic interest and the potential number ofoocytes available [29]. Similar experimentswere developed in mice, also includingother cell types (fetal ovary and testis,immature Sertoli cells, spleen, thymus,

312 C. Thibault

macrophages) from three strains of mice[15]. The overall conclusion is that culturedfetal cells performed better with respect topregnancy initiation and young born thanadult cells with the exception of cumuluscells, which produced the highest overallpregnancy and living young in cattle as wellas in mice.

Forsberg et al. [30] have shown that thecell type that combines extended prolifera-tion in culture and a high pregnancy initia-tion and calving rates, is fetal genital ridgecells (GR). Cultured GR cells derived frompresumptive PGC cells from 40–80 day oldbovine fetuses, may double as many as85 times overall in 407 days of culture andup to 75 times after dilution to single-cellculture. After 68 doubling, the BF15c3 cellline maintained a normal caryotype. CultureGR cells used in nuclear transfer initiatedpregnancies in 40% of recipient heifers(197) and 9% produced living calves(Tab. VIII).

The variability in performance amongdifferent GR cell lines was high; one linegave the highest rate of calving (15.2% ofembryos transferred). Although this rate isnot better than with that adult cumulus cells,the advantage of GR cells is their broad abil-ity to proliferate in culture that makes themsuitable to transfection.

3.7. Cell cycle stageof the transferred nucleus

In very few papers the transferred nu-cleus was really at a precise stage of the cellcycle. As shown by Zou et al. [31] the per-centage of goat fibroblasts in G0 + G1 var-ied according to the type of culture. It was61.6% in normal culture medium, 71.7% af-ter a 3-day starvation in serum-deprivedmedium, and 74.2%, in confluent cultures.They cloned with cells either in G0(G0 + G1?) or in M. They obtained 5 livekids from 32 cloned G0 embryos but nobirth when M (M + G2?) cells weretransferred.

Willmut et al. [32] suggested that induc-tion of nuclear quiescence (stage G0) by se-rum starvation was critical in allowingdonor somatic cells to support the develop-ment of sheep cloned embryos. However,almost simultaneously, Liu et al. [33] ob-tained 3 lambs from morula nuclei in DNAsynthesis or in mitosis providing that theenucleated oocytes were at similar stages.The best result was obtained from the trans-fer of blastomere nuclei in mitosis tometaphase II oocytes.

Ono et al. [16] have shown in mice thateven fetal fibroblasts arrested at themetaphase of the cell cycle were able to


Table VIII. Percentage of calves obtained after transfer of cells from different tissues. Results werehigh with cumulus cells and gonadal ridge cells although variable according to the cell line, and lowwith adult year cells (Forsberg et al. [29]).

Cell typesNo. ofembryotransfers

No. ofpregnant

cows

% ofinitiation

No. ofcalves

% calvesper

transfer

Gonadal ridge cells(GR)

Fetal body cells fromthe same embryo asGR

Adult ear cells

Adult cumulus cells

197

48

142

34

78

23

48

21

39.6

47.9

33.8

61.8

17

3

7

5

9.3–15.2

9.1

5.1

15.2

initiate development up to birth. A similarconclusion has been drawn by Zhou et al.[34]. They showed that ES cells inmetaphase provided a higher developmentrate to morula (71.8 to 86.7%) or blastocyststages (26.8 to 70%) than ES cellinterphasic nuclei (16.4 to 28.2%) and (5.9to 14.1%) respectively. Implantation siteswere higher and 2 pups were born. ES cellsin metaphase were also used in the study ofgene expression in cloned mice (seeSuemizu et al. [18]).

Vignon et al. [35] obtained a similar per-centage of blastocysts from fibroblast nucleieither after starvation or in proliferation andcalves were also regularly obtained what-ever the cell cycle stage.

Hill et al. [36], studying fetal loss duringthe first trimester of pregnancy, observedsimilar pregnancy rates for recipient cowscarrying embryos from serum-fed or serumstarved nuclei, either at day 30 of pregnancy(46%, 25/54 and 44%, 29/66 respectively)or at day 90 (4/10 and 6/10).

With pure G1 fibroblasts either from45-day old bovine fetuses or 4–15 year old

adult cattle, Kasinathan et al. [37] obtained17 to 67% blastocysts. They transferredpure G1-embryos and also G0-embryos into50 recipients for each cell-cycle type. Theembryonic losses were higher with G0 em-bryos and finally only G1 embryos gavebirth to calves. They also observed that se-rum starvation was not as efficient as con-fluence to obtain “pure” G0 nucleus cells.

Thus the postulate that the G0 cell stageis the condition to obtain cloned offspring,is not valid.

3.8. Abnormal gene expression

3.8.1. Embryo

Daniels et al. [38] compared the tran-scription of six developmentally importantgenes (octamer-binding transcription fac-tor 4, (Oct-4)); fibroblast growth factors 2and 4, (FGF2 and FGF4); fibroblast growthfactor receptor 2 (FGFr2); interleukin 6,(IL6); GP 130 and polymerase A) in clonedbovine embryos (granulosa cells) and inin vitro fertilized embryos.

314 C. Thibault

Table IX. Transcription of 7 genes in morulas-blastocysts from fertilized or cloned bovine embryos(Daniels [38]).

Gene transcript In vitro fertilized embryos Cloned embryos

Morula Blastocyst Morula Blastocyst

Oct-4 6/6 6/6 5/5 5/5

IL-6 4/4 5/5 1/4 4/4

FGF 2 4/4 2/4 2/3 3/3

FGF 4 6/7 10/10 0/2 2/7

FGF r2 7/9 9/10 2/6 3/6

Gp 130 5/5 5/5 2/2 3/3

Poly A 16/16 16/16 6/6 11/11

Morula-blastocyst stage embryos de-rived from nuclear transfer proceduresshowed abnormal transcription of IL6,FGF-4 and FGFr2 as shown in Table IX.Transcripts of Oct-4, IL-6, FGF-4, FGFr2are not readily detectable in granulosa cells,however, reprogramming of the somaticcell nucleus is sufficient in NT embryos toproduce the correct pattern of embryonicOct-4 transcription factor. The absence ofIL-6 transcripts in 3 of the 4 NT morulae isindicative of a delayed onset of IL-6 tran-scription in NT embryos. There is also anabsence of FGF-4 transcripts in the majorityof NT morula - blastocysts. In addition,FGFr2 transcripts are significantly lower inNT embryos. The lack or delayed expres-sion of IL-6, FGF-4 and FGFr2 detected inNT embryos is likely to have significantconsequences on the developmental viabil-ity of NT embryos. It has been previouslyshown that mouse embryos that are homo-zygous for targeted disruptions of theFGF-4 (Feldman et al. [39]) or FGFr2(Arman et al. [40]) genes, die shortly afterimplantation although they are able to in-duce uterine response (decidualization).

A more thorough investigation of Oct-4expression in mouse clones showed anoma-lies on its distribution [41]. IVF embryosexpressed Oct-4 after the 4-cell stage, and atthe blastocyst stage. This expression ismaintained in the inner cell mass (ICM) butdown-regulated in the trophectoderm (TE).IVF blastocysts, cultured on the feeder-celllayer, form outgrowths containing Oct-4expressing cells.

In cumulus cell clones, Oct-4 is ex-pressed at the correct stage (> 4-cell stage)but in the majority of the blastocysts, spatialexpression is incorrect and maintained inthe trophectoderm (55% vs. 4–13%). Allcloned outgrowths lack Oct-4-mRNA-expressing cells.

Abnormal Oct-4 expression may be as-sociated with aberrant expression of othercrucial developmental genes, and Oct-4

levels alone account for the majority of thefailures currently observed for somatic cellcloning.

3.8.2. Placenta

Humpherys et al. [42] studied, by micro-array analysis on RNA, the expression ofmore than 10 000 genes in the placentas ofneonatal mice and of neonatal cloned micederived from nuclear transfer (NT) of eithercultured embryonic stem cells (ES) orfreshly isolated cumulus cells (Fig. 1).

Using the same technical approach,Suemizu et al. [18] compared the expressionof approximately 12 000 unique mousegenes in the placentas of ES-cloned miceand of mice from one-cell donor nuclei, ascontrols2. They described the following fiveabnormal events:– the expression of some imprinted genes

was severely depressed (H19, 10-fold)while IGF2 and IGF2r expression washigher (4.8- and 3.2-fold respectively);

– there was altered expression of regulatorygenes such as DNA methyltransferaseDnmt-1, Dmnt3a and Dnmt3b, that wereupregulated (3.8-fold); similarly, expres-sion of histone acetyl transferase washigher in ES-derived NT placentas thanin control placentas;

– there was increased expression of onco-genes and tumor associated proteingenes;

– this was overexpression of the gene in-volved in placental growth, such asPlac-1. Its sustained expression late ingestation and its ectopic expression inspongiotrophoblast cells could contributeto placentomegaly (see Sect. 3.2.1);

– many novel genes were identified, suchas Pitrm-1 such a new member of themetalloprotease family. This gene, nor-mally expressed in actively remodelingtissues, such as bone, is abnormallly reg-ulated in ES-derived NT placenta.


2 ES cell nuclei are pluripotent, and the nuclei of fer-tilized eggs are totipotent.

This impressive list of defective geneexpression detected by microarray studies,brings to light the necessity of the following:– to more thoroughly explore the control

of gene expression during normal dev-elopment;

– to study if these genes are correctly regu-lated in apparently healthy living clonedanimals.

In contrast to ES cell clones, it seems thatthe changes in expression of the genesIGFII, IGFIIr and H19 do not occur in pla-centas of clones that are produced from fetaland adult fibroblast cells (Ogura et al. [17]).

3.9. Imprinting defaults

Imprinted genes are erased in the primor-dial germ cell (PGC) before their gameticpatterns are re-established during gameto-genesis according to the sex of individuals.

Lee et al. [43] studied the developmentof cloned mouse embryos produced fromday 11.5 to day 13.5 PGC. An erasure pro-cess occurred between day 10.5 and day11.5. Cloned embryos produced from day12.5 to day 13.5 PGC, showed growth retar-dation and embryonic lethality around day9.5 of pregnancy. Imprinted genes becamebiallelic or silent. Cloned embryos pro-duced from day 11.5 PGC survived longer,at least until day 11.5. Several intermediate

states of genomic imprinting were seen inthese embryos as well as the demethylationstatus. This means that imprinted genes arenot erased simultaneously and that somedifferences exist between embryos.

4. ABNORMALITIES INDUCEDBY IN VITRO MANIPULATIONAND CULTURE

Some abnormalities described in clonedembryos, placenta and fetuses or the youngare relevant to in vitro manipulation andculture. It is interesting to note that cumuluscells which give the highest percentage ofsuccess are used without culture. Amongstanomalies observed after cloning with cul-tured adult cells, the following must bementionned: telomere shortening, abnormalmethylation, aberrant gene expression,large offspring syndrome, and heteroploidy.

4.1. Telomere shortening

During in vitro culture of sheep fibro-blasts from adult tissue, the telomeres short-ened by 157 bp per population doubling[24]. A similar regression of 100 bp perpopulation doubling occurred during theculture of fibroblasts from a 13-year-olddairy cow [26].

316 C. Thibault

cumulus 232 cumulus 54

ES 188 ES 33

84 148 40 34 2013

A B

Figure 1. The number of genes that show a 2-fold difference in average expression and differing sig-nificantly from controls (P < 0.05). Shadowed areas indicated gene expression equally modified incumulus- and ES cell-derived clones. (A) gene expression reduced in NT placentas; (B) gene expres-sion elevated in NT placentas (from Humpherys et al. [42]).

4.2. Abnormal methylation

Shi and Haaf [44] showed that the inci-dence of abnormal methylation patterns in2-cell mouse embryos (loss of methylationof one or both nuclei, completely methyl-ated nuclei) differs between in vivo and invitro fertilized embryos. Methylated DNAwas visualized by a well-characterizedmonoclonal antibody against 5-methyl cy-tosine (MeC) counterstained with DAPI.When in vitro fertilized eggs were culturedup to the blastocyst stage, the percentage ofabnormal methylation varied according tothe culture media and the strain (Tab. X).

They found a good correlation betweenthe number of abnormally methylatedmouse 2-cell embryos and the embryonicloss during the preimplantation period.Moreover, 95% of the embryos from nonsuperovulated females developed in cultureto the blastocyst stage versus only 86% ofthose from superovulated females. These

percentages correlated with the abnormalmethylation rates, respectively 10 and 20%of the 2-cell embryos.

4.3. Lower or delayed geneexpression

Blastocyst formation, pregnancy ratesand birth rates are lower than in vivo whenoocyte maturation, fertilization and earlydevelopment occurs in vitro. In order toidentify what step is responsible, Knijn et al.[45] studied in bovine embryos, the expres-sion of six marker genes important in devel-opment: plakophilin and desmocollin 2 ofembryonic origin, glucose transporter 1,E-cadherin, heat shock protein 70-1, poly Apolymerase of both maternal and embryonicorigin.

Blastocysts were produced in vitro fromoocytes from different follicle sizes: (1)3–8 mm follicles, (2) preovulatory folliclesbefore LH surge and (3) preovulatory


Table X. Development in vitro of eggs fertilized either in vivo (VVF) or in vitro (IVF) according to theculture medium and the percentages of DNA abnormal methylation (Shi and Haaf [44]).

Medium

In vitro development Methylated DNA (MeC staining)

No. analysed % developingto blastocyst

No. analysed % with abnormalmethylation

pattern

(1) IVF mHam F10*

(1) IVF mTC199**

(1) IVF M 16 c***

(1) IVF M 16

(2) IVF M16

20

28

29

45

56

20

75

83

80

72

7

22

12

18

35

71

32

9

22

40

VVF oocytes from:

Normal females

Superovulated females

40

50

95

86

100

30

10

20(1) Strain B6C3F1; (2) strain NMRI.* mHam F10 (modified); ** mTC199 (modified); *** M16 c = M16 + BSA, dithiothreitol and heparin.

follicles 24 h after the LH surge. Oocytesfrom groups 1 and 2 were matured in vitro,whereas oocytes from group 3 had under-gone their maturation in vivo (4). The con-trol group was composed of blastocystsdeveloped entirely in vivo.

Oocyte maturation in vitro did not mod-ify gene expression. However, fertilizationand in vitro culture up to the blastocyst stageinduced alterations in the relative abun-dance of 3 transcripts: expression glucosetransporter 1 that was significantly higher(P < 0.05) for in vivo blastocysts comparedwith in vitro blastocysts; desmocollin 2 andplakophilin tended to be higher (P < 0.1)whereas no differences were found for heatshock protein 70-1 (see [46]), E-cadherinand poly(A) polymerase.

Thus alterations in the relative abun-dance of three transcripts in blastocysts pro-duced in vitro cannot primarily be attributedto the origin of the oocyte, but are likely tohave been induced by post-maturation orfertilization and culture conditions, or both.

Bertolini et al. [46] investigated the ef-fect of the embryo production system ongrowth and transcription rate on day 7 andday 16 bovine embryos. In vitro embryoswere produced after in vitro maturation ofthe oocytes in modified TCM 199, fertiliza-tion based on the Parish technology, andco-culture with bovine oviductal epithelialcells.

Thirty-four day-7 in vivo (control group)and twenty-eight day-7 in vitro producedembryos (IVP) were non surgically trans-ferred to female recipients synchronous tothe donor (± 12 h). In the control group,18 were transferred as single embryos and16 were transferred as dual embryos, each inone uterine; from IVP embryos, 14 weretransferred as single and 14 as dual embryos.

A wide variability in development andelongation was observed in both groups,however, retarded development, predomi-nantly at the embryonic disc level, was evi-dent in the IVP concepti. The embryonic

disc tended to be smaller in female IVP thanthat of control group female embryos. Inci-dentally, it must be mentioned that the em-bryonic disc was larger and the embryolonger in male than in female concepti,irrespective to the group, indicating awell-known fast-growing trend for the maleembryos.

The significance of genes examined inthis study resides in their close associationwith embryo growth, development and rec-ognition of pregnancy. Embryonic tran-scripts of IGF-I and IGF-II, their receptorsIGF-Ir and IGF-IIr, glucose transporters 1and 3 (GlT-1, -3), and interferon τ (inf τ)were determined by real-time quantitativePCR.

Day 7 IVP embryos presented lower to-tal mRNA levels than the control (P < 0.05)but this difference was generally reduced onday 16.

As reported previously by Knijn et al.[45], Day 7 IVP embryos had significantlysmaller amounts of transcripts of IGF com-ponents (except IGF-I), of glucose trans-porter-1, transporter-3 and IFN τ than controlembryos.

In conclusion, the in vitro production ofbovine embryos negatively affects gene ex-pression on day 7 and the rate of develop-ment on day 16.

4.4. Large offspring syndrome

The large offspring syndrome (LOS) isobserved in bovine, ovine and murine off-spring following transfer of in vitro-pro-duced embryos or cloned embryos, and ischaracterized by a multitude of pathologicalchanges, of which extended gestation lengthand increased birth weight are predominantfeatures.

It has been clearly shown that LOS iscaused by culture conditions during in vitroprocedures. Two steps might be involved:fertilization or culture.

318 C. Thibault

Young et al. [47] collected in vivo ovinezygotes and cultured them in vitro for 5 daysbefore their transfer in recipient ewes. Outof 48 fetuses recovered at day 125 of preg-nancy, 12 weighed more than 5 kg with amaximum weight of 8.2 kg. In these LOSfetuses, expression of IGF-IIr was reducedby 30–60% relative to the control group andthere was a complete loss of methylation inthe methylated region of IGF-IIr. Thesefacts suggest that the preimplantationembryo may be vulnerable in culture toepigenetic alterations of imprinted genes.

Sinclair et al. [48] studied the influenceof a 5 day culture of in vivo fertilized eggson the growth of ovine fetuses at day 61 andday 125 of pregnancy. Fetuses derived from

co-cultured embryos were 14% heavier thancontrol fetuses (P < 0.01) by day 61 and34% heavier (P < 0.001) by day 125 ofpregnancy. When embryos were cultured inSOF medium, increased weight was lowerbut remained significant (P < 0.01). Thusthe LOS did not result from in vitro fertiliza-tion, but of the culture of the zygote.

Lazzari et al. [49] analyzed birth weight,and the differences in the following parame-ters of day 7 and day 12 bovine blastocysts,according to their origin in vivo or in vitro/in vivo:– the number of cells in day 7 embryos

(Tab. XI);– the birth weight of calves derived from in

vivo and in vitro production (Tab. XII);


Table XI. Number of cells in day 7 bovine embryos after a five day culture (Lazzari et al. [49]).

Culture mediaor in vivo

No. of oocytes No. of day 7blastocysts

Blastocysts /oocytes (%)

No. cells /blastocyst

SOF-bsa

SOF-serum

Sheep oviduct

830

1105

848

162

200

180

19.5 ± 3.1

18.1 ± 4.5

21.6 ± 2.4

183 ± 37

179 ± 23

142 ± 34**P < 0.01.

Table XII. Gestation length and birth weight in cattle. Abnormal body weight at birth after in vitroculture (Lazzari et al. [49]).

Group No. calves Gestation lengthdays

Birth weightkg

SOF-bSA

SOF-serum

Early developmentin sheep oviduct

In vivo fertilizationand development(superovulated cows)

In vivo artificialinsemination

23

10

34

63

24

281.1 ± 7.1*

280.1 ± 5.9*

279.2 ± 5.3*

279.4 ± 5.1*

281.7 ± 4.3*

52.8 ± 9.4*

56.7 ± 12.1*

44.1 ± 5.5**

41.1 ± 3**

43.4 ± 4.3**

* P < 0.01; ** P < 0.05.

– the relative abundance, in day 12 pre-implantation bovine embryos, of thetranscripts of genes developmentally im-portant, glucose transporter 1, 3 and 4(GlT-1, -2, -3, -4), Cu/Zn superoxydedismutase (SOD), histone H4.1, bFGF,IGF-Ir and IGF-IIr (Tab. XIII).

As shown in Table XI, the percentage ofblastocysts was similar whatever the type ofculture, however, the number of cells perblastocyst was significantly higher inin vitro culture systems. When the threetypes of blastocysts were transferred in therecipient cows and developed to term, theduration of gestation was similar in allgroups, but birth weights were significantlyhigher when embryos had been cultured(Tab. XII).

Out of the 9 genes studied, 5 showed asignificantly higher level of transcription inin vitro cultured embryos than after earlydevelopment in the sheep oviduct (Tab. XIII).

Reduced IGF2r protein in mice andsheep is assumed to stimulate growth and tobe responsible for LOS. However, the work

of Chavatte-Palmer et al. [22] suggests an-other explanation. In cloned and controlcalves showing a highly significant differ-ence of body weight at birth (55.1 ± 2.7 kgvs. 43.7 ± 0.5 – 45.7 ± 1.5 kg, P < 0.001),they studied the levels of thyroxin, IGF-I,IGF-II, cortisol and leptin from birth to7 days. One of the main endocrine differ-ences between cloned and control calvesconcerns the high level of leptin (P < 0.01)in LOS calves. In humans, Koistinen et al.[50], Schubring et al. [51], Matsuda et al.[52], Ong et al. [53], Varvarigou et al. [54]have shown that the concentration of leptinin human cord blood is closely correlated tobirth weight and placental weight and thatleptin is of placental origin. These observa-tions suggest that LOS calves could resultfrom a hyper secretion of leptin by theplacenta.

4.5. Heteroploidy, mixoploidy

It is well-known that heteroploidy in-creases the risk of defective developmentand abortion. Is this chromosome anomaly

320 C. Thibault

Table XIII. Abundance of transcripts of 9 genes in 12-day bovine embryos (Lazzari et al. [49]).

Transcripts SOF-bSA SOF-serum Sheep oviduct In vivo

Hsp 70.1 + +P < 0.05

normal control

SOD +P < 0.05

+P < 0.05

normal control

Glut-1 normal normal normal control

Glut-3 +P < 0.05

+P < 0.05

+P < 0.05

control

Glut-4 +P < 0.05

(+) normal control

IGF1-R +P < 0.05

+P < 0.05

normal control

IGF2-R normal normal normal control

BFG +P < 0.05

normal normal control

Histone-4 normal normal normal control

increased by in vitro fertilization or/and invitro culture?

Viuff et al. [55–59] studied the ploidy ofin vivo and in vitro produced bovine em-bryos. In both groups, mixoploid embryoswere present, and consisted mainly of dip-loid-triploid cells. Mixoploidy occurredmore frequently than polyploidy. From day2 to day 5 after fertilization of in vivomatured oocytes, there was an increase(+5–31%) in the percentage of mixoploidcells. The frequency of polyploidy andmixoploidy of day 3 embryos derived fromin vivo matured oocytes was lower than inembryos from in vitro matured oocytes (re-spectively 1% versus 6% and 16% versus25%).

They followed the evolution of hetero-ploidy in embryos from in vitro maturedoocytes at day 7–8 and day 12 post-fertil-ization. At day 7–8, some polyploid cellswere present in 96% of the trophectoderm(TE) and in 58% of the embryonic disc(ED). An average 10.8 ± 2.2% and 7.7 ±2.2% polyploid cells were found in TE andED respectively. Embryos were transferredto the uteri on day 7 and then recovered onday 12. An average of 7.3 ± 4.4% and 3.1 ±2.6% polyploid cells was found in TE andED, respectively. Thus there was a decreasein the frequency of polyploid cells from day7–8 to day 12. Unfortunately the levels ofheteroploidy on day 7–8 and day 12 in em-bryos from in vivo matured oocytes werenot given.

Booth et al. [58] studied the level ofploidy in day 7 NT blastocysts. Amongst the112 blastocysts that were analyzed, 26(23.2%) possessed completely diploid nucleiand 6 (5.4%) contained only pure polyploidnuclei. The others were mixoploids and thetetraploid state was the predominant condi-tion (85.5%).

An inverse relationship between blasto-cyst total cell number and total percentageof chromosome abnormalities is observedwithin embryos.

In conclusion, there is a high percentageof chromosome number abnormalities in NTembryos and this percentage is not derivedfrom cumulus donor cells (0.23 ± 0.12). Invitro maturation of oocytes may increase thepercentage of heteroploid cells in concepti,mainly in the trophectoderm.

These observations suggest that the highlevel of ploidy defects contributes to highand prolonged embryonic losses of nucleartransfer embryos.

V. CONCLUSION

This review deals first with the morpho-logical, cellular and genetic abnormalitiesobserved during the development of clonedembryos, fetuses and placenta, and second,what abnormalities are relevant to cloningitself or to the in vitro manipulation. Thefollowing conclusions can be drawn.

Abnormal gene expression and methyl-ation, which affect fetuses and placentas aswell result in placental hypotrophy or hy-pertrophy, in errors in imprinting, epigeneticreprogrammation, and in X inactivation.

With a somatic nucleus at any stage ofthe cell cycle, cloning is possible providedthe recipient egg cytoplasm is at the samestage.

The culture after egg reconstruction andbefore the transfer to the recipient female,leads to lower or delayed gene expression,to epigenetic modifications and to an in-crease in the frequency of polyploidy ormixoploidy.

Two approaches have been used to studygene expression: either individual genesknown to be important in development suchas Oct-4, or thousands of genes by micro-analysis on RNA. Oct-4 is overexpressedand its spatial distribution is incorrect. Withmicro array, clearly important changes havebeen observed such as over expression ofIGF2 / IGF2r, DNMT-1, Dnmt 3a and 3b. Inthe placenta, the specific growth factor


Plac-1 is also over expressed. These resultsare somewhat disappointing because aroundmore than one hundred genes are either overor lower expressed in the placenta of livingmouse offspring. Apparently these genesare not necessary to fetal development. Ithas been recently shown in the bacteria, Ba-cillus subtilis, that only 271 of its 4100genes permit its survival and growth [59].Thus the interpretation of gene misexpressionto determine causes of abnormal develop-ment of cloned embryos, fetuses and pla-centas needs the knowledge, first, of thegene expression in in vivo produced em-bryos, and second, the evolution of thesegenes during growth, puberty and reproduc-tion in the surviving cloned offspring.

REFERENCES

[1] Pace MM, Augenstein ML, Betthauser M et al.Ontogeny of cloned cattle to lactation. BiolReprod 2002, 67: 334–339.

[2] Tamashiro KL, Wakayama T, Akutsu UH,Yamazaki Y, Lachey JL, Wortman MD, SeeleyRJ, D’Alessio DA, Woods SC, Yanagimachi R,Sakai RR. Cloned mice have an obese phenotypenot transmitted to their offspring. Nature Med2002, 8: 262–267.

[3] Heyman Y, Chavatte-Palmer P, Le Bourhis D,Camous S, Vignon X, Renard JP. Frequency andoccurrence of late-gestation losses from cattlecloned embryos. Biol Reprod 2002, 66: 6–13.

[4] Renard JP, Zhou QI, Le Bourhis D,Chavatte-Palmer P, Hue I, Heyman Y, Vignon X.Nuclear transfer technologies: between suc-cesses and doubts. Theriogenology 2002, 57:203–222.

|5] Gao S, Gasparini B, McGarry M, Ferrier T,Harkness L, De Souza P, Wilmutt Y. Germinalvesicle material is essential for nucleus remodel-ling after nuclear transfer. Biol Reprod 2002, 67:928–934.

|6] Tani T, Kato Y, Tsunoda Y. Direct exposure ofchromosomes to non activated ovum cytoplasmis effective for bovine somatic cell nucleus. BiolReprod 2001, 64: 324–330.

[7] Kim JM, Ogura A, Nagata M, Aoki F. Analysisof the mechanism for chromatin remodelling inembryo reconstructed by somatic nuclear trans-fer. Biol Reprod 67, 2002: 760–766.

[8] Loi P, Clinton M, Barboni B, Fulka J, CappaiI P,Feil R, Moor RM, Pyak G. Nuclei of nonviable

ovine somatic cells develop into lamb after nu-clear transplantation. Biol Reprod 2002, 6:126–132.

[9] Reik W, Dean W, Walter J. Epigenetic repro-gramming in mammalien development. Science2001, 293: 1089–1093.

[10] Kierszenbaum AL. Genomic imprinting andepigenetic reprogramming: unearthing the gar-den of forking paths. Mol Reprod Dev 2002, 63:269–272.

|11] Hata K, Okano M, Lei H, Li E. Dnmt3L cooper-ates with the Dnmt3 family of de novo DNAmethyltransferases to established maternal im-prints in mice. Development 2002, 129:1983–1993.

[12] Cezar GG, Bartolomei MS, Forsberg EJ, FirstNL, Bishop MD, Eilertsen KJ. Genome-wide al-terations in cloned bovine fetuses. Biol Reprod2003, 68: 1009–1014.

[13] Rybouchkin A, Heindryckx B, Van der Elst J,Dhont M. Developmental potential of clonedmouse embryos reconstructed by a conventionaltechnique of nuclear injection. Reproduction2002, 124: 197–207.

[14] Koo DB, Kang YK, Choi HE, Sun Paark J, KIMHN, Oh KB, SON DS, Park H, Lee KK, HanYM. Aberrant allocations of inner cell mass andtrophectoderm cells in bovine nuclear transferblastocysts. Biol Reprod 2002, 67: 487–492.

[15] Wakayama T, Yanagimachi R. Mouse cloningwith nucleus donor cells of different age andtype. Mol Reprod Dev 2001, 58: 376–383.

[16] Ono Y, Shimozawa N, Ito M, Kono T. Clonedmice from fetus fibroblast cells arrested atmetaphase by a serial nuclear transfer. BiolReprod 2001, 64: 44–50.

[17] Ogura A, Inoue K, Ogonuki N, Lee J, Kohda T,Ishino F. Phenotypic effects of somatic cellscloning in the mouse. Clon Stem Cells 2002, 4:397–405.

[18] Suemizu H, Aiba K, Yoshikawa T, Sharow AA,Shimozawa N, Tamaoki N, Ko MS. Expressionprofiling of placentomegaly associated with nu-clear transplantation of mouse ES cells. Dev Biol2003, 253: 36–53.

[19] Tanaka S, Oda M, Toyoshima Y, Wakayama T,Tanaka M, Yoshida N, Hattori N, Ohgane J,Yanagimachi R, Shiota K. Placentomegaly incloned mouse concepti caused by expansion ofthe spongiotrophoblast layer. Biol Reprod 2001,65: 1813–1821.

[20] Inoue K, Khoda T, Lee J, Ogonuki N, MochidaK, Noguchi Y, Tenamura K, Kaneko-Ishino T,Ishino F, Ogura A. Faithful expression of im-printed genes in cloned mice. Science 2002, 295:297.

[21] Hill JR, Schlafer DH, Fisher PJ, Davies CJ. Ab-normal expression of trophoblast major

322 C. Thibault

histocompatibility complex class I antigens incloned bovine pregnancies is associated with apronounced endometrial lymphocytic response.Biol Reprod 2002, 67: 55–63.

[22] Chavatte-Palmer P, Heyman Y, Richard C,Monget P, Le Bourhis D, Kann G,Chillard Y,Vignon X, Renard J. Clinical, hormonal andhematologic characteristics of bovine calves de-rived from nuclei from somatic cells. BiolReprod 2002, 66: 1596–1603.

[23] Renard JP, Chastant S, Chesné P, Richard C,Marchal J, Cordonnier N, Chavatte P, Vignon X.Lymphoid hupoplasia and somatic cloning. TheLancet 1999, 353: 1789–1491.

[24] Shiels PG, Kind AJ, Campbell KHS,Waddington D, Wilmutt Y, Colman A, SchniekeAE. Analysis of telomere lengths in clonedsheep. Nature 1999, 399: 316–317.

[25] Lanza RP, Cibelli JB, Blackwell C, CristopholoVJ, Francis MK, Baerlocher GM, Mak J,Scheritzer M, Chavez EA, Sawyer N, LandsdorpPM, West MD. Extension of cell-life-span andtelomere length in animals cloned from senes-cent somatic cells. Science 2000, 288: 665–669.

[26] Tian XC, Xu J, Yang X. Normal telomere lengthfound in cloned cattle. Nature Genet 2000, 26:272–273.

[27] Miyashita N, Shiga K, Yonai M, Kaneyama K,Kobayashi S, Kojima T, Goto Y, Khisi M, Aso H,Suzuki T, Sakaguchi M, Nagai T. Remarkabledifferences in telomere lengths among clonedcattle derived from different cell types. BiolReprod 2002, 66: 1649–1655.

[28] Xue F, Tian XC, Du F, Kubota C, Tanega M,Dinnyes A, Dai A, Levine H, Pereira LV, YangX. Aberrant patterns of X-chromosome inactiva-tion in bovine clones. Nature Genet 2002, 31:216–220.

[29] Kato Y, Tani T, Tsunoda Y. Cloning of calvesfrom various somatic cell types of male and fe-male adult, newborn and fetal cows. J ReprodFertil 2000, 120: 231–237.

[30] Forsberg EJ, Stelchenko NS, Augenstein ML,Betthauser JM, Childs LA, Eilertsen KJ et al.Production of cloned cattle from in vitro system.Biol Reprod 2002, 67: 327–333.

[31] Zou X, Wang Y, Cheng Y, Yang Y, Ju H, Tang H,Shen Y, Mu Z, Xu S, Du M. Generation ofcloned goats (Capra hircus) from transfectedfœtal fibroblast cells, the effect of donor cell cy-cle. Mol Reprod Dev 2002, 61: 164–172.

[32] Wilmut I, Schnieke A, McWhir J, Kind AJ,Campbell KHS. Viable offspring derived fromfetal and adult mammalian cells. Nature 1997,385: 810–813.

[33] Liu L, Dal Y, Moor RM. Nuclear transfer insheep embryos: the effect of cell cycle coordina-tion between nucleus and cytoplasm and the use

of in vitro matured oocytes. Mol Reprod Dev1997, 47: 255–264.

[34] Zhou Q, Jouneau A, Brochard V, Adenot P,Renard JP. Developmental potential of mouseembryo reconstructed from metaphase embry-onic stem cell nuclei. Biol Reprod 2001, 65:412–419.

[35] Vignon X, Chesné P, Le Bourhis D, Fléchon JE,Heyman Y, Renard JP. Developmental potentialof bovine embryos reconstructed fromenucleated matured oocytes fused with culturesomatic cells. C-R Acad Sci Paris, Sci Vie, 1998,321: 735–745.

[36] Hill JR, Burghardt RC, Jones K, Long CR,Looney CR, Shin T, Spenver TE, Thompson JA,Winger QA, Westhusin ME. Evidence for pla-cental abnormality as the major cause of mortal-ity in first trimester somatic cell cloned bovinefetuses. Biol Reprod 2000, 63: 1787–1794.

[37] Kasinathan P, Knott JG, Wang Z, Jerry DJ, RoblJM. Viable offspring derived from fetal and adultmammalian cells. Nature Biotech 2001, 19:1176–1178.

[38] Daniels R, Hall V, Trouson AO. Analysis of genetranscription in bovine nuclear transfer embryosreconstructed with granulosa cell nuclei. BiolReprod 2000, 63: 1034–1040.

[39] Feldman B, Poueymirou W, Papaioannou VE,DeChiara TM, Goldfarb M. Requirement ofFGF-4 for post-implantation mouse develop-ment. Science 1995, 267: 246–249.

[40] Arman E, Haffner-Krauz R, Chen Y, Heath JK,Lonai P. Targeted disruption of fibroblast growthfactor (FGF) receptor 2 suggests a role for FGFsignaling in pre-gastrulation mammalian devel-opment. PNAS, USA, 1998, 95: 5082–5087.

[41] Boiani M, Eckardt S, Schöller HR, McLaughlinKJ. Oct 4 distribution and level in mouse clones:consequences for pluripotency. Genes Dev 2002,16: 1209–219.

[42] Humpherys D, Eggan K, Akutsu H, Friedman A,Hochedlinger K, Yanagimachi R, Lander ES,Golub TR, Jaenisch R. Abnormal gene expres-sion in cloned mice derived from embryonicstem cell and cumulus cell nuclei. Proc NatlAcad Sci USA, 2002, 99: 12889–12894.

[43] Lee J, Inoue K, Ono R, Ogonuki N, Khoda T,Kaneko-Ishino T, Ogura A, Ishino F. Erasinggenomic imprinting memory in mouse clonedembryos produced from day 121.5 primordialgerm cells. Development 2002, 129: 1807–1817.

[44] Shi W, Haaf T. Aberrant methylation patterns atthe two-cell stage as an indicator of early devel-opmental failure. Mol Reprod Dev 2002, 63:329–334.

[45] Knijn HM, Wrenzycki C, Hendriksen PJM, VosPLAM, Hermann D, Van der Weijden GC,Nieman H, Dieleman SJ. Effects of oocyte


maturation regimen on the relative abundance ofgene transcripts in bovine blastocysts derived invitro or in vivo. Reproduction 2002, 124:365–375.

[46] Bertolini M, Beam SW, Shim H, Bertolini LR,Moyer AL Famula TR, Anderson GB. Growth,development and gene expression by in vivo-and in vitro-produced day 7 and day 16 bovineembryos. Mol Reprod Dev 2002, 63: 318–328.

[47] Young LE, Fernandez K, McEvoy TG,Butterwith SC, Guteriez CG, Carolan C,Broadbent PJ, Robinson JJ, Wilmut I, SinclairKD. Epigenetic change in IGF-2R is associatedwith fetal overgrowth after sheep embryo cul-ture. Nature Genet 2001, 27: 153–154.

[48] Sinclair KD, McEvoy TG, Maxfield EK, MaltinCA, Young LE, Wilmut I, Broadbent PJ, Robin-son JJ. Aberrant fetal growth and developmentafter in vitro culture of sheep zygotes. J ReprodFertil 1999, 116: 177–186.

[49] Lazzari G, Wrenzycki C, Hermann D et al. Cel-lular and molecular deviation in bovine inviitro-produced embryos are related to the largeoffspring syndrome. Biol Reprod 2002, 67:767–775.

[50] Koistinen HA, Koivisto VA, Andersson S,Karonen SL, Kontula K, Oksanen L, TeramoKA. Leptin concentration in cord blood corre-lates with intrauterine growth. J Clin EndocrinolMetab 1997, 82: 3328–3330.

[51] Schubring C, Kiess W, Englaro P, Rascher W,Dötsch J, Hanitsch S, Attanasio A, Blum WF.Level of leptin in maternal serum, amniotic fluid,and arterial and venous cord blood: relation toneonatal and placental weight. J Clin EndocrinolMetab 1997, 82: 1480–1483.

[52] Matsuda J, Yokota I, Ida M, Murikami T, NaitoE, Ito M, Shima K, Kuroda Y. Serum leptin

concentration in cord blood: relationship to birthweight and gender. J Clin Endocrinol Metab1997, 82: 1642–1644.

[53] Ong Ken KL, Ahmed ML, Sherriff A, Woods KA, Watts A, Golding J, Dunger DB. Cord bloodleptin is associated with size at birth and predictsinfancy weight gain in humans. J ClinEndocrinol Metab 1999, 84: 1145–1148.

[54] Varvarigou A, Manzoros CS, Beratis NG. Cordblood leptin concentrtions in relation tointrauterine growth. Clin Endocrinol 1999, 50:177–183.

[55] Viuff D, Greve T, Avery B, Hyttel P, BrockhoffPB, Thomsen PD. Chromosome aberrations inin vitro produced bovine embryos at days 2–5post-insemination. Biol Reprod 2000, 63:1143–1148.

[56] Viuff D, Hendriksen PJM, Vos PLAM ,Dieleman SJ, Bibby BM, Greve T, Hyttel P,Thomsen PD. Chromosomal abnormalities anddevelopmental kinetics in the in vivo-developedcattle embryos at day 2 to 5 after ovulation. BiolReprod 2001, 65: 204–208.

[57] Viuff D, Palsgaard A, Rickords L, Lawson LG,Greve T, Schmidt M, Avery B, Hyttel P, Thom-sen PD. Bovine embryos contain a higher pro-portion of polyploid cells in the trophectodermthan in the embryonic disc. Mol Reprod Dev2002, 62: 483–488.

[58] Booth PJ, Viuff D, Shijian T, Holm P, Greve T,Callesen H. Numerical chromosome errors inday 7 somatic nuclear transfer bovineblastocysts. Biol Reprod 2003, 68: 922–928.

[59] Kobayashi K, Ehrlich SD, Albertini A, Amati G,Andersen KK, AsaiK, Ashikaga S, Aymerich S.Essential bacillus subtilis gene. Proc Natl AcadSci USA 2003, 100: 4678–4683.

To access this journal online:www.edpsciences.org

324 C. Thibault

charles thibault - edp sciences

Documents