Granulocyte–macrophage colony-stimulating factor promotes human ...

3 downloads 0 Views 406KB Size Report
The cytokine granulocyte–macrophage colony-stimulating factor (GM-CSF) is synthesized in the female reproductive tract and has been implicated in the growth ...
Human Reproduction vol.14 no.12 pp.3069–3076, 1999

Granulocyte–macrophage colony-stimulating factor promotes human blastocyst development in vitro

C.Sjo¨blom1,3, M.Wikland1 and S.A.Robertson2 1Fertilitetscentrum,

402 29 Go¨teborg, Sweden and 2Department of Obstetrics and Gynaecology and Reproductive Medicine Unit, University of Adelaide, Adelaide 5005, Australia

3To

whom correspondence should be addressed

The cytokine granulocyte–macrophage colony-stimulating factor (GM-CSF) is synthesized in the female reproductive tract and has been implicated in the growth and development of the preimplantation embryo in rodent and livestock species. To examine the effect of GM-CSF on human embryo development in vitro, surplus frozen 2–4-cell embryos were cultured in media supplemented with 2 ng/ ml recombinant human GM-CSF. The addition of cytokine increased the proportion of embryos that developed to the blastocyst stage from 30 to 76%. The developmental competence of these blastocysts, as assessed by hatching and attachment to extracellular matrix-coated culture dishes, was also improved by GM-CSF. The period in culture required for 50% of the total number of blastocysts to form was reduced by 14 h, and blastocysts grown in GM-CSF were found to contain ~35% more cells, due primarily to an increase in the size of the inner cell mass. The beneficial effect of GM-CSF was exerted in each of two sequential media systems (IVF-50/S2 and G1.2/G2.2) and was independent of the formulation of recombinant cytokine that was used. These data indicate that GM-CSF may have a physiological role in promoting the development of the human embryo as it traverses the reproductive tract in vivo, and suggest that addition of this cytokine to embryo culture media may improve the yield of implantationcompetent blastocysts in human in-vitro fertilization programmes. Key words: blastocyst/culture/embryo/GM-CSF/growth factor

Introduction A major focus of human embryo research is to develop an invitro culture system that more closely replicates the female reproductive tract and thereby is able to support the routine growth of embryos to the blastocyst stage. In human in-vitro fertilization (IVF), culture and transfer of blastocysts is likely to improve implantation and pregnancy success by providing a means for identifying the most robust embryos for transfer to patients (Me´ne´zo et al., 1992; Olivennes et al., 1994; Gardner and Lane, 1997, 1998; Jones et al., 1998). Through allowing the selection of smaller numbers of more competent embryos, blastocyst transfer may help to reduce the frequency © European Society of Human Reproduction and Embryology

of multiple births resulting from IVF. Transfer of blastocysts to the uterus is clearly also a more physiological procedure than transfer at early cleavage stages in that better synchronization between the embryo and the developmental status of the endometrial tissue is achieved. Studies in rodents and livestock species suggest that the growth and development of the preimplantation embryo is regulated by an array of cytokines and growth factors secreted from epithelial cells lining the oviduct and uterus (Pampfer et al., 1991; Robertson et al., 1994). Their synthesis occurs in precise, differentially regulated spatial and temporal patterns driven predominantly by ovarian steroid hormones but also by factors in seminal plasma (Robertson et al., 1996; Tremellen et al., 1998). Embryos of many species, including human, have been shown to express receptors for many of the growth factors secreted by the tract (Pampfer et al., 1991; Sharkey et al., 1995), and the addition of growth factors and cytokines to embryo culture medium has been shown to promote blastocyst development in the human (Dunglison et al., 1996; Lighten et al., 1998; Martin et al., 1998) and in a number of mammalian species (reviewed by Kane et al., 1997). Granulocyte–macrophage colony-stimulating factor (GMCSF) is a cytokine originally identified as a product of activated T-lymphocytes involved in the proliferation and differentiation of myeloid haematopoietic cells (Ruef and Coleman, 1990). GM-CSF is produced by oestrogen-primed epithelial cells in the oviduct and uterus in mice (Robertson et al., 1992), sheep (Imakawa et al., 1993) and women (Zhao and Chegini, 1994; Giacomini et al., 1995). Animal studies suggest that GM-CSF can act as a survival factor for the developing embryo. Murine preimplantation embryos express the α-chain of the GM-CSF receptor, and culture in recombinant GM-CSF-containing media has beneficial effects on murine embryo development to blastocyst stage and subsequently on their capacity to hatch from the zona pellucida and attach to the culture dish (Robertson et al., 1991, 2000). Recent studies have shown that genetically GM-CSFdeficient mice have retarded blastocyst formation with a significantly lower number of blastomeres, principally due to a diminished inner cell mass size. The effects on the null mice also include decreased fetal size and increased rates of fetal resorption during late gestation and mortality during early post-natal life (Robertson et al., 1999). The development of in-vitro produced bovine embryos is improved by addition of GM-CSF to culture media (de Moraes and Hansen, 1997). Exposure of ovine embryos to GM-CSF in vitro increases their implantation potential through enhanced expression of the antiluteotrophic signal interferon (IFN)-τ (oTP-1) in trophectoderm cells (Imakawa et al., 1993) . 3069

C.Sjo¨ blom, M.Wikland and S.A.Robertson

Together these findings suggest that GM-CSF is a potential regulator of human preimplantation embryo development. The purpose of this study was to explore this possibility by investigating the effects of recombinant human GM-CSF on human embryo development in vitro.

Table I. Classification of 2–4-cell embryos Embryo grade

Morphology

A

Regular blastomeres without fragments Regular or irregular blastomeres, 艋30% fragments Regular or irregular blastomeres, ⬎30% fragments 50% of the blastomeres dead after thawing

B

Materials and methods The embryos used in this study were donated by couples undergoing IVF treatment at Fertilitetscentrum AB, Go¨ teborg, Sweden. For culture experiments, embryos frozen at the 2–4-cell stage were thawed at or beyond their 1 year storage limit in liquid nitrogen. The blastocysts used for the differential staining experiment were cultured from excess embryos, surplus to treatment and freezing requirements. Ethics approval for the current study was obtained from the research ethics committee at University of Go¨ teborg (number 700-96) Media and embryo culture All media were from Scandinavian IVF Science AB, Go¨ teborg, Sweden. Embryo culture was performed in IVF-50, S2, G1.2 and G2.2, and micro-drops were overlaid with Ovoil-150. Gamete-100, comprised of HEPES-buffered IVF-50, was used for differential staining. IVF-50 is a low glucose and low phosphate medium modified from the original human tubal fluid (HTF) formulation (Quinn et al., 1985). Media S2, G1.2 and G2.2 are modifications of the original G1 and G2 (Gardner, 1994; Barnes et al., 1995). Modifications to the original culture media formulations include reduced glutamine, EDTA and phosphate concentrations, the inclusion of specific vitamins in G2.2, and the use of human serum albumin rather than bovine serum albumin. S2 medium is a modification of the G2 medium, supplemented with human serum albumin. Apart from the insulin in S2, the media do not contain any protein or growth factors other than pharmaceutical grade human serum albumin. The precise formulations of these media are withheld by Scandinavian IVF Science for commercial reasons. Ovarian stimulation and in-vitro fertilization Patients received 300 µg buserelin gonadotrophin-releasing hormone agonist (GnRHa, Suprecur; Hoechst, Frankfurt, Germany) three times daily intranasally, starting 1 week before expected menses and lasting for 2 weeks. Down-regulation was confirmed by a serum oestradiol content of ⬍0.2 nmol/l. Patients were then given recombinant follicle stimulating hormone (r-FSH; Gonal-F; Serono Laboratories, Aubonne, Switzerland; 150–225 IU/day s.c.). The starting dose was dependent on the patient’s age and/or previous response during ovarian stimulation (Wikland et al., 1994). The ovarian response was monitored by ultrasound and serum oestradiol concentration as previously described (Bergh et al., 1997). GnRHa and rFSH were administered until there was at least one follicle ⬎18 mm in mean diameter and two others 艌16 mm. Finally, oocyte maturation was triggered by one s.c. injection of 10 000 IU of human chorionic gonadotrophin (HCG, Profasi; Serono Laboratories). Oocytes were retrieved 36–38 h after HCG administration, assessed morphologically and fertilized in vitro. The embryos were cultured in IVF-50 and frozen on day 2 using a three-step propanediol cryopreservation kit (Freeze Kit 1, Scandinavian IVF Science) according to the manufacturer’s instructions. Recombinant GM-CSF Two different commercial sources of recombinant human (rh)GMCSF were used in these experiments. A laboratory grade preparation was obtained from R&D Systems Europe Ltd, Abingdon, Oxon, UK,

3070

C D

Table II. Distribution of grades among thawed 2–4-cell embryos allocated to control and test media Embryo grade

Control (%)

GM-CSF (%)

A B C D n

20 22 32 26 50

20 27 33 20 49

Grades are defined in Table I. GM-CSF ⫽ granulocyte–macrophage colony-stimulating factor.

and a pharmaceutical grade preparation, Molgramostim (Leucomax) was obtained from Schering & Plough, Madison, NJ, USA. The biological activity of both recombinant cytokine preparations was measured in a bioassay employing a GM-CSF responsive cell line (human myeloid TF-1 cell line), essentially as previously described (Kitamura et al., 1989). Duplicate serial 1:2 dilutions of both preparations were incubated with 2000 TF-1 cells in 200 µl of RPMI1640 (GIBCO/BRL, Grand Island, NY, USA) supplemented with 10% fetal calf serum (FCS; Commonwealth Serum Laboratories, Parkville, Victoria, Australia), 5⫻10–5 mol/l β-mercaptoethanol and antibiotics. After 2 days, cultures were pulsed with 1 µCi of [3H]thymidine (Amersham, Arlington Heights, IL, USA) for 6 h, harvested onto glass fibre paper using a PHD automated cell harvester (Cambridge Technology Inc., Cambridge, MA, USA) and radioactivity measured in a liquid scintillation beta counter. Embryo thawing, allocation and culture Frozen 2–4-cell embryos were thawed in four steps using a propanediol method for embryo thawing (Thaw Kit 1, Scandinavian IVF Science) following instructions given by the manufacturer. The viable embryos were classified and graded according to criteria listed in Table I. To avoid bias the embryos were randomly allocated, with regard to patient and embryo grade, into the different culture groups (Table II). The embryos were cultured in groups of five embryos per drop. To avoid the toxic effects of ammonium, released due to metabolism and breakdown of amino acids, the culture medium was renewed every 48 h until hatching occurred. In two experiments the embryos were cultured in 20 µl drops of IVF-50 containing 2 ng/ml rhGM-CSF (R&D Systems; diluted 1:25 000 from stock material) or carrier [2 ng/ml bovine serum albumin (BSA), diluted 1:1000 from stock material]. Culture drops were covered by 4 ml Ovoil-150 in Falcon 3004 dishes (Becton-Dickinson Labware, Franklin Lakes, NJ, USA). When blastocysts were observed these were transferred into 1 ml of S2 medium in Falcon 3037 dishes, supplemented with 5% fetal calf serum (FCS), plus 2 ng/ml rhGMCSF or carrier. Developmental stage was scored every 8 h from thawing until 2300 on day 8 (200 h post-insemination). In a third experiment the embryos were transferred from IVF-50

GM-CSF promotes human blastocyst development

Table III. Rate and extent of embryo development in the presence or absence of recombinant human granulocyte–macrophage colony-stimulating factor (rhGM-CSF) Experiment

1 2 3 Total

Control

rhGM-CSF

n

Bφ (%)

H (%)

A (%)

n

Bφ (%)

H (%)

A (%)

16 16 18 50

6 6 3 15

3 3 1 7

0 0 0 0

15 16 18 49

9 13 15 37

8 13 8 29

5 7 4 16

(38) (38) (17) (30)a

(50) (50) (33) (47)b

(60) (81) (83) (76)a

(89) (100) (53) (78)b

(62) (54) (50) (55)

Bφ ⫽ number of viable thawed 2–4-cells reaching blastocyst stage (aP ⬍ 0.0001); H ⫽ number of blastocysts which fully or partially hatch (bP ⫽ 0.009); A ⫽ number of blastocysts which attached to the culture dish.

into S2 medium at the 6–8 cell stage. Additions of GM-CSF and carrier were the same as in the two previous experiments. When blastocysts were observed they were transferred to Falcon 3037 dishes, coated 24 h previously with Biomatrix EHS (Boehringer Ingelheim Bioproducts, Heidelberg, Germany). Developmental stage was scored every 8 h from thawing until 2300 h on day 8 (200 h post-insemination). In a fourth experiment, embryo culture was performed in two different sequential media systems using two different commercial sources of rhGM-CSF. After thawing, the embryos were cultured first in G1.2 or IVF-50. At 6–8-cell stage the embryos were transferred into G2.2 or S2. The experiment included six groups: (i) G1.2/G2.2 alone; (ii) G1.2/G2.2 containing 2 ng/ml rhGM-CSF (R&D Systems); (iii) G1.2/G2.2 containing 2 ng/ml Molgramostim (diluted 1:75 000 from stock material); (iv) IVF-50/S2 alone; (v) IVF-50/S2 containing 2 ng/ml rhGM-CSF (R&D Systems); (vi) IVF-50/S2 containing 2 ng/ml Molgramostim. Developmental rate was scored every eighth hour until expanded blastocyst stage. Blastocysts were scored on day 5 at 120 h post-insemination according to criteria described previously (Dokras et al., 1993). Briefly, grade A blastocysts exhibited an expanded cavity with a distinct trophectoderm (TE) and an eccentrically located inner cell mass (ICM); grade B blastocysts were not yet expanded but otherwise morphologically identical to A; and grade C blastocysts exhibited poor morphology characterized by a number of degenerative foci in the ICM and TE and a poorly developed blastocoel cavity. Embryo scoring in each of the experiments was performed by the same person (C.S.). Statistical analysis was performed using Fisher’s exact test and independent samples t-test (StatSoft, Inc., Tulsa, OK, USA). Differences in data were considered significant when P ⬍ 0.05. Differential labelling of blastocysts Differential labelling was performed using a modification of a protocol described previously (Handyside and Hunter, 1984). Human blastocysts were cultured from excess embryos, surplus to treatment and freezing. On day 5 of culture (120–124 h post-insemination) the zona was removed in acid Tyrode’s solution containing 4 mg/ml polyvinylpyrrolidone (PVP; 360 000 mol. wt) and embryos were washed once in Gamete-100 and three times in albumin-free S2 containing 4 mg/ml PVP (S2-PVP). The blastocysts were incubated in trinitro-benzene sulphonic acid (TNBS, Sigma Chemical Co., St Louis, MO, USA; 10 mmol/l in S2-PVP pH 8.5, 4°C/20 min in the dark) and washed three times in Gamete-100. TNBS-treated blastocysts were incubated in anti-dinitrophenyl antibody (anti-DNP; Sigma, 0.2 mg/ml diluted in Gamete-100; 37°C/30 min). Embryos were then washed and incubated in guinea-pig complement serum (Sigma; diluted 1:10 in Gamete-100; 37°C/30 min). Embryos were washed again and labelled with fluorochromes (Sigma; 0.05 mmol/l bisbenzim-

ide and 10 µg/ml propidium iodide in Gamete-100, 37°C/30 min). After extensive washing embryos were fixed briefly in 1% paraformaldehyde and 0.5% glutaraldehyde in PBS, mounted under coverslips in 20% glycerol in PBS and examined by fluorescence microscopy using a 400 nm excitation filter. Nuclei stained pink were scored as lysed trophectoderm cells (TE) and blue nuclei were scored as viable inner cell mass cells (ICM).

Results The effect of rhGM-CSF on human preimplantation embryo development in vitro To determine the effect of rhGM-CSF on the in-vitro development of human embryos, thawed 2–4-cell embryos were cultured in the presence of 2 ng/ml rhGM-CSF or carrier. In the first two experiments, 5% FCS was included in the culture media from blastocyst stage onwards, to provide a source of fibronectin and other attachment factors and allow embryo attachment and outgrowth to be assessed. In the third experiment FCS was omitted and blastocysts were transferred to dishes coated with Biomatrix-EHS. In each of the three separate experiments GM-CSF was found to have a strong beneficial effect on embryo development in vitro. The proportion of 2–4-cell embryos developing to the blastocyst and hatching blastocyst stages was significantly increased by the addition of rhGM-CSF to culture medium [37/49 and 15/50 2–4-cell embryos developed to blastocyst in GM-CSF and control media respectively (P ⬍ 0.0001) and 29/49 and 7/50 initiated hatching in GM-CSF and control media respectively (P ⬍ 0.009; Table III)]. Blastocysts grown in GM-CSF (16/37), but not in control media (0/15), attached to the culture dish and showed evidence of trophoblast outgrowth during the ensuing 24 h period (Figure 1). Statistical comparison of the combined data from experiments 1 and 2 (culture media containing 5% FCS from day 5) with experiment 3 (serum-free media) showed that there were no significant differences in the proportion of embryos which developed into blastocysts between the two groups, indicating that the effect of GM-CSF was independent of FCS. Hatching, attachment and trophectoderm outgrowth appeared somewhat diminished in blastocysts grown on EHS-coated dishes compared with those in FCS, an effect which appeared to be due to an embryotoxic influence of the Biomatrix-EHS preparation. 3071

C.Sjo¨ blom, M.Wikland and S.A.Robertson

Figure 1. The effect of recombinant human granulocyte–macrophage colony-stimulating factor (rhGM-CSF) on the rate of development of embryos to the blastocyst stage: (A) an early blastocyst (day 5, 112 h post-insemination) from the control group; (B) an expanded blastocyst (day 5, 112 h post-insemination) cultured in rhGM-CSF; (C) a fully hatched blastocyst attached to the culture dish (day 6, 144 h postinsemination); (D) an attached blastocyst cultured in rhGM-CSF showing trophectoderm outgrowth (arrow; day 8, 200 h post-insemination). Original magnification ⫽ A,B, ⫻ 200; C,D, ⫻ 100.

Figure 2 illustrates the effect of rhGM-CSF on blastocyst formation according to embryo grade at thawing. GM-CSF exerted a comparable effect in all grades of 2–4-cell embryos, with similar increases in the proportion of poor quality compared with good quality embryos forming blastocysts. The majority of embryos cultured in media alone were lost at the 4–16-cell stage (Figure 3). The beneficial effect of rhGM-CSF on blastocyst development appeared to result from rescue of this loss, with an 80% increase in embryos reaching the morula stage. Embryos cultured in the presence of rhGM-CSF had a significantly higher cleavage rate with the first blastocysts forming early on day 4, more than 24 h earlier than the first blastocysts grown in media alone (Figures 1 and 4). Fifty per cent blastocyst development was achieved 14 h earlier in the rhGM-CSF groups compared with the control groups. Figure 2. The effect of human recombinant granulocyte– macrophage colony-stimulating factor (rhGM-CSF) on the development of embryos to blastocysts, according to embryo grade (defined in Table I). Data are the number of embryos developed to blastocyst from experiments 1, 2 and 3 combined, expressed as a percentage of the initial number of cleaved (2–4-cell) embryos. The number of embryos in each group is given in parentheses. Data were analysed by Fisher’s exact test. *Significantly different from corresponding control group: P ⬍ 0.04.

3072

The effect of culture media and source of recombinant cytokine on GM-CSF stimulated blastocyst development To determine if the effect of GM-CSF on human embryos was dependent on the culture media and source of recombinant cytokine, embryos were cultured in two different sequential culture media systems, in the presence of either laboratory grade or pharmaceutical grade rhGM-CSF. Both cytokine

GM-CSF promotes human blastocyst development

Figure 3. The effect of recombinant human granulocyte–macrophage colony-stimulating factor (rhGM-CSF) on the development of embryos to blastocyst, hatching and attachment stages. Data are the number of embryos developed to or beyond each stage, from experiments 1, 2 and 3 combined, expressed as a percentage of the initial number of cleaved (2–4-cell) embryos. 2-C ⫽ 2-cell embryos; 8C ⫽ 8-cell embryos; M ⫽ morula; B ⫽ blastocyst; exp B ⫽ expanded blastocyst; H ⫽ hatching; A ⫽ attached with trophectoderm outgrowth. The number of embryos in each group is given in parentheses. Data were analysed by Fisher’s exact test. *P ⬍ 0.004; **P ⬍ 0.001 denote significant difference from control group.

Table IV. The effect of culture media and source of recombinant cytokine on recombinant human granulocyte–macrophage colony-stimulating factor (rhGM-CSF) stimulated blastocyst development

G1.2/G2.2 alone G1.2/G2.2 ⫹ rhGM-CSF (R&D) G1.2/G2.2 ⫹ Molgramostim IVF-50/S2 alone IVF-50/S2 ⫹ rhGM-CSF (R&D) IVF-50/S2 ⫹ Molgramostim

n

Bφ (%)

A/B/C (%)

23 21 19 38 38 20

7 (30) 15 (71)** 12 (63)* 14 (37) 30 (79)*** 13 (65)*

57/29/14 67/20/13 67/17/17 57/29/14 67/26/7 70/15/15

*P ⬍ 0.05, **P ⬍ 0.01, ***P ⬍ 0.001 denote significant differences from corresponding sequential media alone. Bφ ⫽ number of viable thawed 2–4-cell embryos reaching blastocyst stage; A/B/C ⫽ blastocyst grading (see Materials and methods).

Figure 4. The effect of recombinant human granulocyte– macrophage colony-stimulating factor (rhGM-CSF) on the rate of development of embryos to the blastocyst stage. Data are the number of blastocysts at each time point, from experiments 1, 2 and 3 combined, expressed as a percentage of the total number of blastocysts at 144 h post-insemination. Data were analysed by linear regression analysis (Sigma Plot, Jandel Scientific Software, Corte Madera, CA, USA) and 50% blastocyst formation was found to occur at 108 h and 122 h post-insemination for GM-CSF and control treated embryos respectively. Data were analysed by Fisher’s exact test, and the highest significance was achieved at 122 h post-insemination (P ⬍ 0.0002).

formulations were found to have equivalent bioactivities in the TF-1 cell proliferation assay (data not shown). Neither FCS nor Biomatrix-EHS were used in the embryo culture experiment, so blastocyst attachment was not determined. There were no significant differences between the blastulation rates achieved in the two different culture media systems (Table IV). Both the rate and extent of development of 2–4cell embryos to blastocysts was significantly increased by the

addition of 2 ng/ml rh GM-CSF. The effect was comparable in extent in both G1.2/G2.2 and IVF-50/S2 sequential media combinations. Furthermore, the improvement in blastocyst development was achieved irrespective of the source of recombinant cytokine. The results also show that although culture in rhGM-CSF gives rise to more blastocysts, the distribution in morphological grade was comparable in treatment and control groups (Table IV). The effect of culture in GM-CSF on blastomere number and allocation. To investigate the effect of culture with GM-CSF on cell number and allocation to inner cell mass and trophectoderm cell lineage, blastocysts cultured with and without rhGM-CSF were analysed by immunosurgery and differential staining. Blastocysts cultured in the presence of rhGM-CSF had a significantly higher total cell number compared with blastocysts cultured in media alone (Figure 5). An increase in the number of trophectoderm cells, and particularly in the number of inner cell mass cells, each contributed to the greater cell number in GM-CSF-stimulated blastocysts. 3073

C.Sjo¨ blom, M.Wikland and S.A.Robertson

Figure 5. The effect of recombinant human granulocyte– macrophage colony-stimulating factor (rhGM-CSF) on the number of total cells (TCN), inner cell mass (ICM) cells and trophectoderm (TE) cells in day 5 blastocysts (120–124 h post-insemination). Values are mean ⫾ SD of blastocysts cultured in 2 ng/ml rhGMCSF (n ⫽ 11) and blastocysts cultured in media alone (n ⫽ 10). Data were analysed by independent samples t-test. *P ⬍ 0.03 and **P ⬍ 0.0001 denote significantly different from corresponding control.

Discussion GM-CSF is expressed during early pregnancy in the reproductive tract of all mammalian species so far studied, and experiments in mice and livestock animals suggest that this cytokine may have a physiological role in promoting the development of the preimplantation embryo (Robertson et al., 1991, 2000; Imakawa et al., 1993; de Moraes and Hansen, 1997). The experiments reported in the current study show that addition of rhGM-CSF to human embryo culture media can dramatically increase the proportion of frozen–thawed 2–4-cell embryos that proceed to the blastocyst, hatching and attachment stages of development. The beneficial effect is exerted on poor quality as well as good quality frozen–thawed 2–4-cell embryos. The effect of GM-CSF is independent of the formulation of recombinant cytokine preparation and occurred in both of two different sequential culture media systems. Moreover, frozen– thawed embryos cultured in the presence of rhGM-CSF have a significantly increased cleavage rate which is known to be a sign of high developmental potential (McKiernan and Bavister, 1994). This was also reflected in experiments where differential staining was employed to measure the total number and allocation of blastomeres. Blastocysts cultured in the presence of GM-CSF were found to have a significantly increased total cell number, with an increased proportion of cells located in the inner cell mass, than blastocysts cultured for the same period of time in media alone. Morphological assessment of the blastocysts yielded in culture showed that despite giving rise to more blastocysts in a shorter time period, GM-CSF did not alter the range of qualities seen in blastocysts cultured in vitro. These results concur with previous findings in rodents, where increased developmental potential afforded to embryos by GM-CSF is associated with promotion of inner cell mass viability. Blastocysts recovered from GM-CSF null mice have 3074

a decreased total cell number and inner cell mass size compared to blastocysts from wild-type littermates, and culture of embryos in the presence of recombinant murine GM-CSF can alleviate the effect of GM-CSF deficiency (Robertson et al., 2000). These data together with the findings in this study suggest that GM-CSF promotes blastomere survival and/or proliferation, and that the effect is preferentially exerted in the inner cell mass. Although the molecular events that culminate in these effects are not fully understood, our recent experiments using mouse blastocysts show that GM-CSF can promote glucose uptake through binding to trophectoderm cells via the α-chain of the GM-CSF receptor (Robertson et al., 1999). Increased glucose uptake is known to stimulate metabolic activity, and might therefore promote cell division. It has been shown in other cells that cytokine withdrawal can induce apoptosis (Hetts, 1998) and the diminished survival rate of embryos cultured in growth factor-free medium could be due to apoptosis resulting from metabolic starvation. Our recent findings (unpublished data) have shown that human blastocysts express the α-chain but not the β-chain of the heterodimeric GM-CSF receptor, but whether improved glucose transport is stimulated by GM-CSF in human blastocysts remains to be determined. The ability of GM-CSF to preserve inner cell mass viability is important since at least in other species the number of inner cell mass cells in the blastocyst at the time of implantation appears to be a key parameter influencing implantation success, and subsequent fetal size and health. Depletion of inner cell mass cells is characteristic of blastocysts generated in vitro, and this defect is correlated with abnormal fetal size late in gestation (Leese et al., 1999). In mice, the beneficial effect of insulin in in-vitro culture is mediated through its capacity to promote the survival and/or proliferation of inner cell mass cells (Kaye et al., 1992), whereas cytokines with embryotoxic properties such as tumour necrosis factor (TNF)α appear to constrain embryo development by inhibiting the survival and/ or proliferation of inner cell mass cells (Pampfer et al., 1995). A likely role for GM-CSF in the physiological development of the preimplantation embryo in the human reproductive tract is suggested by the temporal pattern of expression of GMCSF in the oviduct and uterus, which coincides with the time of fertilization and early embryo development and implantation. Expression of GM-CSF in the human Fallopian tube has been shown to be dependent on the stage of the menstrual cycle, with peak expression occurring during the mid–late proliferative and early–mid secretory phases (Zhao and Chegini, 1994). A similar pattern of cycle-related synthesis of GM-CSF occurs in the uterus, where GM-CSF secreted principally from endometrial epithelial cells is present throughout the cycle with a moderate increase during the late proliferative and early secretory phases (Giacomini et al., 1995). The presence of a large repertoire of different cytokines and growth factors in the reproductive tract during the time of preimplantation embryo development and implantation indicates that many different growth regulators may act in concert to orchestrate optimal embryo development. Other factors known to be expressed in the female reproductive tract have been reported to promote human embryo development.

GM-CSF promotes human blastocyst development

Addition of leukaemia inhibitory factor (LIF) to a complex serum-free human embryo culture media increased the blastulation rate from 18 to 44% (Dunglison et al., 1996). The effect of LIF was limited to increased blastocyst formation, since no effects beyond blastulation were observed. Culture of human embryos in the presence of heparin-binding epidermal growth factor (HB-EGF) also improves the proportion of embryos developing through to the blastocyst stage as well as their developmental competence as assessed by hatching, adherence to extracellular matrix proteins and trophectoderm outgrowth (Martin et al., 1998). The extent of the effect was comparable to that reported for GM-CSF in the current study, with blastulation rates increased from 41% in the control group to 71% in the presence of HB-EGF. In contrast to the GMCSF, however, HB-EGF did not improve the developmental rate or cell number and its effect was limited to good quality embryos. Addition of insulin-like growth factor (IGF)-1 to invitro culture media has also been shown to benefit human embryo development (Lighten et al., 1998), increasing the blastulation rate from 35% in the control to 60% in the treatment group. The effect of IGF-I also included an increase in blastocyst cell number, due entirely to an increase of cells in the inner cell mass. It is not known whether the mechanisms underlying the effects of all of these growth factors are related, but it could be speculated that each has influence on the metabolic activity of inner cell mass and/or trophectoderm cells, leading to an improved rate of cell division and/or protection from apoptosis. In-vitro culture conditions for the culture of human embryos are generally considered to be suboptimal and this is believed to compromise the quality of embryos and may contribute to the high rates of implantation failure seen in human IVF. Culture to the blastocyst stage provides a means for selecting the most developmentally competent embryos for transfer to patients as well as achieving a better synchronization between the embryo stage and uterine development. Thus, the transfer of blastocysts to the uterus may lead to higher implantation rates and help to reduce the number of multiple births resulting from IVF. However, clinical application of these findings in human IVF must await the results of further experiments designed to elucidate the mechanism of action of GM-CSF in the embryo, and to evaluate the effect of culture with GMCSF on the subsequent in-vitro developmental potential of blastocysts in animal models. Finally, any transfer to patients of blastocysts cultured in GM-CSF or other growth factors will be undertaken with caution, given that the events of early preimplantation development are now recognized to have longterm consequences for the health of the individual in childhood and in adult life (Seamark and Robinson, 1995). Acknowledgements This study was supported by a Serono Nordic AB Research Scholarship in Assisted Reproduction (C.S.). S.A.R. is the recipient of an NH&MRC R Douglas Wright Research Award.

References Barnes, F.L., Crombie, A., Gardner, D.K. et al. (1995) Blastocyst development and birth after in-vitro maturation of human primary oocytes,

intracytoplasmic sperm injection and assisted hatching. Hum. Reprod., 10, 3243–3247. Bergh, C., Howles, C.M., Borg, K. et al. (1997) Recombinant human follicle stimulating hormone (r-hFSH; Gonal-F) versus highly purified urinary FSH (Metrodin HP): results of a randomized comparative study in women undergoing assisted reproductive techniques. Hum. Reprod., 12, 2133–2139. de Moraes, A.A. and Hansen, P.J. (1997) Granulocyte-macrophage colonystimulating factor promotes development of in vitro produced bovine embryos. Biol. Reprod., 57, 1060–1065. Dokras, A., Sargent, I.L. and Barlow, D.H. (1993) Human blastocyst grading: an indicator of developmental potential? Hum. Reprod., 8, 2119–2127. Dunglison, G.F., Barlow, D.H. and Sargent, I.L. (1996) Leukaemia inhibitory factor significantly enhances the blastocyst formation rates of human embryos cultured in serum-free medium. Hum. Reprod., 11, 191–196. Gardner, D.K. (1994) Mammalian embryo culture in the absence of serum or somatic cell support. Cell Biol. Int., 18, 1163–1179. Gardner, D.K. and Lane, M. (1997) Culture and selection of viable blastocysts: a feasible proposition for human IVF? Hum. Reprod. Update, 3, 367–382. Gardner, D.K. and Lane, M. (1998) Culture of viable human blastocysts in defined sequential serum-free media. Hum. Reprod., 13 (Suppl. 3), 148–159. Giacomini, G., Tabibzadeh, S.S., Satyaswaroop, P.G. et al. (1995) Epithelial cells are the major source of biologically active granulocyte macrophage colony-stimulating factor in human endometrium. Hum. Reprod., 10, 3259–3263. Handyside, A.H. and Hunter, S. (1984) A rapid procedure for visualising the inner cell mass and trophectoderm nuclei of mouse blastocysts in situ using polynucleotide-specific fluorochromes. J. Exp. Zool., 231, 429–434. Herrler, A., Krusche, C.A. and Beier, H.M. (1998) Insulin and insulinlike growth factor-I promote rabbit blastocyst development and prevent apoptosis. Biol. Reprod., 59, 1302–1310. Hetts, S.W. (1998) To die or not to die: an overview of apoptosis and its role in disease [see comments]. J. Am. Med. Assoc., 279, 300–307. Imakawa, K., Helmer, S.D., Nephew, K.P. et al. (1993) A novel role for GMCSF: enhancement of pregnancy specific interferon production, ovine trophoblast protein-1. Endocrinology, 132, 1869–1871. Jones, G.M., Trounson, A.O., Gardner, D.K. et al. (1998) Evolution of a culture protocol for successful blastocyst development and pregnancy. Hum. Reprod., 13, 169–177. Kane, M.T., Morgan, P.M. and Coonan, C. (1997) Peptide growth factors and preimplantation development. Hum. Reprod. Update, 3, 137–157. Kaye, P.L., Bell, K.L., Beebe, L.F. et al. (1992) Insulin and the insulin-like growth factors (IGFs) in preimplantation development. Reprod. Fertil. Dev., 4, 373–386. Kitamura, T., Tange, T., Terasawa, T. et al. (1989) Establishment and characterization of a unique human cell line that proliferates dependantly on GM-CSF, IL-3, or erythropoietin. J. Cell. Physiol., 140, 323–334. Leese, H.J., Donnay, I. and Thompson, J.G. (1999) Human assisted conception: a cautionary tale. Lessons from domestic animals. Hum. Reprod., 13 (Suppl. 13), 184–202. Lighten, A.D., Moore, G.E., Winston, R.M. et al. (1998) Routine addition of human insulin-like growth factor-I ligand could benefit clinical in-vitro fertilization culture. Hum. Reprod., 13, 3144–3150. Martin, K.L., Barlow, D.H. and Sargent, I.L. (1998) Heparin-binding epidermal growth factor significantly improves human blastocyst development and hatching in serum-free medium. Hum. Reprod., 13, 1645–1652. McKiernan, S.H. and Bavister, B.D. (1994) Timing of development is a critical parameter for predicting successful embryogenesis. Hum. Reprod., 9, 2123–2129. Me´ ne´ zo, Y., Hazout, A., Dumont, M. et al. (1992) Coculture of embryos on Vero cells and transfer of blastocysts in humans. Hum. Reprod., 7 (Suppl. 1), 101–106. Olivennes, F., Hazout, A., Lelaidier, C. et al. (1994) Four indications for embryo transfer at the blastocyst stage. Hum. Reprod., 9, 2367–2373. Pampfer, S., Arceci, R.J. and Pollard, J.W. (1991) Role of colony stimulating factor-1 (CSF-1) and other lympho-hematopoietic growth factors in mouse pre-implantation development. Bioessays, 13, 535–540. Pampfer, S., Vanderheyden, I., Vesela, J. et al. (1995) Neutralization of tumor necrosis factor alpha (TNF alpha) action on cell proliferation in rat blastocysts by antisense oligodeoxyribonucleotides directed against TNF alpha p60 receptor. Biol. Reprod., 52, 1316–1326. Quinn, P., Kerin, J.F. and Warnes, G.M. (1985) Improved pregnancy rate in human in vitro fertilization with the use of a medium based on the composition of human tubal fluid. Fertil. Steril., 44, 493–498. Robertson, S.A., Lavranos, T.C. and Seamark, R.F. (1991) In vitro models of

3075

C.Sjo¨ blom, M.Wikland and S.A.Robertson the maternal fetal interface. In Wegmann, T.G., Nisbett-Brown, E. and Gill, T.G. (eds) The Molecular and Cellular Immunobiology of the MaternalFetal Interface. Oxford University Press, New York, pp. 191–206. Robertson, S.A., Mayrhofer, G. and Seamark, R.F. (1992) Uterine epithelial cells synthesize granulocyte-macrophage colony-stimulating factor and interleukin-6 in pregnant and nonpregnant mice. Biol. Reprod., 46, 1069– 1079. Robertson, S.A., Seamark, R.F., Guilbert, L.J. et al. (1994) The role of cytokines in gestation. Crit. Rev. Immunol., 14, 239–292. Robertson, S.A., Mayrhofer, G. and Seamark, R.F. (1996) Ovarian steroid hormones regulate granulocyte-macrophage colony-stimulating factor synthesis by uterine epithelial cells in the mouse. Biol. Reprod., 54, 183–196. Robertson, S.A., Roberts, C.T., Farr, K.L. et al. (1999a) Fertility impairment in granulocyte-macrophage colony-stimulating factor-deficient mice. Biol. Reprod., 60, 251–261. Robertson, S.A., Sjo¨ blom, C., Jasper, M.J. et al. (2000) Granulocytemacrophage colony-stimulating factor promotes glucose transport and inner cell mass viability in murine blastocysts (in press). Ruef, C. and Coleman, D.L. (1990) Granulocyte-macrophage colonystimulating factor: pleiotropic cytokine with potential clinical usefulness. Rev. Infect. Dis., 12, 41–62. Seamark, R.F. and Robinson, J.S. (1995) Potential health problems stemming from assisted reproduction programmes. Hum. Reprod., 10, 1321–1322. Sharkey, A.M., Dellow, K., Blayney, M. et al. (1995) Stage-specific expression of cytokine and receptor messenger ribonucleic acids in human preimplantation embryos. Biol. Reprod., 53, 974–981. Tremellen, K.P., Seamark, R.F. and Robertson, S.A. (1998) Seminal transforming growth factor beta1 stimulates granulocyte-macrophage colony-stimulating factor production and inflammatory cell recruitment in the murine uterus. Biol. Reprod., 58, 1217–1225. Wikland, M., Borg, J., Hamberger, L. et al. (1994) Simplification of IVF: minimal monitoring and the use of subcutaneous highly purified FSH administration for ovulation induction. Hum. Reprod., 9, 1430–1436. Zhao, Y. and Chegini, N. (1994) Human fallopian tube expresses granulocytemacrophage colony stimulating factor (GM-CSF) and GM-CSF alpha and beta receptors and contain immunoreactive GM-CSF protein. J. Clin. Endocrinol. Metab., 79, 662–665. Received on June 7, 1999; accepted on September 20, 1999

3076