The HLA-G 14bp gene polymorphism and decidual HLA-G 14bp gene expression in pre-eclamptic and normal pregnancies

2.1. Study groups 

Participants were recruited to the study from the Delivery Ward at St. Olavs Hospital, Trondheim, Norway, from 2002 to 2006. Due to the sampling of decidua basalis tissue, only women undergoing caesarean section (CS) were recruited. Pre-eclamptic cases were defined as persistent blood pressure above 140/90mmHg and proteinuria of more than 0.3g/24h or 2+ or higher according to a dipstick test, developing after 20 weeks of pregnancy (Gifford et al., 2000). Cases (n=31) were analyzed both in total and subgrouped into severe/mild PE (n=26/5) (Sibai et al., 2005) or early-onset/late-onset (n=26/5) PE (clinical manifestations before or after the start of the 34th week of gestation). Controls (n=29) were healthy women with normal pregnancies undergoing CS for various reasons considered irrelevant to the aim of this study, i.e. breech presentation, birth anxiety and previous CS. Control women had no history of pregnancies with PE, recurrent spontaneous abortion (RSA) or fetal growth restriction (FGR). In addition, since maternal blood was not collected from the women first enrolled in the study and to increase the statistical power, 14 non-pregnant women with former normal pregnancies were recruited for maternal genotyping. Pregnancies with chromosomal aberrations, fetal and placental structural abnormalities or suspected perinatal infections were not included. Only singletons were included in the study. Informed consent was obtained from all participants, and the study was approved by the Regional Committee for Medical Research Ethics.

2.2. Biological samples 

Umbilical cord blood (from 31 cases and 29 controls) was stored as serum, and maternal peripheral blood (from 20 cases and 34 controls) was stored in EDTA (one case and 21 controls, including the non-pregnant controls) or as serum (19 cases and 13 controls). Decidua basalis tissue (from 30 cases and 29 controls) was obtained by vacuum aspiration of the placental bed after the placenta was delivered during CS (Staff et al., 1999), and stored in RNA-later (Ambion, Huntington, UK). For quality control, representative pieces of the collected decidual tissue were fixed in 10% neutral-buffered formalin, paraffin embedded and cut in sections at 4μm in a motorized microtome. Presence of EVTs were confirmed by immunohistochemical staining with a mAb against cytokeratin 7 (anti-CK7mAb; clone OV-TL 12/30, DakoCytomation, Glostrup, Denmark) and ChemMate Envision Detection Kit Peroxidase/DAB (DakoCytomation), as previously described (Eide et al., 2007). Only specimens with confirmed presence of EVTs, i.e. decidua basalis, were included for further studies.

2.3. Genotyping of the HLA-G 14bp gene polymorphism 

Genomic DNA was extracted from 200μl serum using the QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany) or from 200μl EDTA–blood with the MagAttract DNA Blood Mini M48 Kit (Qiagen) on the BioRobot M48 Workstation (Qiagen). Genotyping of the HLA-G 14bp polymorphism was performed with HLA-G forward primer 5′-CCTGATGTGTGTGGGTTGTT-3′ and reverse primer 5′-TGAGACAGAGACGGAGACAT-3′. The forward primer was labelled with 6-carboxy fluorescein (6-FAM) in the 5′ (Sigma–Genosys, Haverhill, UK). Forty PCR amplification cycles were done in a Mastercycler®ep (Eppendorf, Hamburg, Germany) and analyzed on an ABI 3130xl Genetic Analyzer (Applied Biosystems, Foster City, CA, USA). Two samples of each genotype were DNA sequenced and verified as HLA-G specific.

2.4. HLA-G 14bp gene expression analysis by RT-PCR 

Decidua basalis tissue samples (80–110mg) were disrupted and homogenized in a rotor–stator homogenizer (Ultra Turax T25, IKA, Stanfen, Germany). Total RNA was extracted by RNeasy Midi Kit (Qiagen), and cDNA prepared from 0.4μg total RNA using the Reverse Transcriptase Core Kit (Eurogentec, Liegè, Belgium). Using the HLA-G RT-PCR primers 5′-GCTCAGATTGAAAAGGAGGGAG-3′ and 5′-GAGACGGAGACATCCCAGCC-3′, 45 cycles of PCR amplification were done in an Applied Biosystems 2720 Thermal Cycler. PCR products were analyzed by gel electrophoresis and UV-detection in a Gel Doc 1000 system (Bio Rad Laboratories, Hercules, CA, USA), and all transcripts from two cases and one control were verified as HLA-G specific by DNA sequencing.

2.5. Statistical analysis 

The clinical data were expressed as mean value with corresponding S.D., and Mann–Whitney U-test or Fisher's exact test was used for comparisons (Table 1). Maternal and fetal HLA-G 14bp genotype frequencies were compared to Hardy–Weinberg expectations using χ2-tests. Allele frequencies, genotype frequencies and gene expression profiles in cases and controls were compared by Fischer's exact test. General linear model was performed to study the relation between the HLA-G 14bp genotype and fetal growth. For comparison of birth weights, the difference between actual birth weight and expected birth weight for the gestational age and sex was calculated from the birth weight curves of Marsal et al. (1996), and then divided by the S.D. to transform the variable (birth weight) to standard score. The standard score is expressed as units of S.D. and used as a measure for birth weight deviation (BWD). Since birth weight normally increases with parity, fetal growth analysis was adjusted for parity. P<0.05 was considered statistically significant.

Table 1. Clinical characteristics

		PE (n=31)	Controls (n=29)a
	Maternal age (years)	30±4	30±5
	Gestational age (weeks)	32±3b	39±1
	Systolic BP (mmHg)	158±14b	121±15
	Diastolic BP (mmHg)	96±10b	72±8
	Birthweight (g)	1558±636b	3499±400
	Primipara	21 (68%)c	9 (31%)

Values are expressed as mean±S.D. Cases are compared to controls. PE, pre-eclampsia; BP, blood pressure.

a Non-pregnant controls (n=14) are not included. b P<0.001; Mann–Whitney test. c P<0.05; Fisher's exact test.

3.1. Patient characteristics 

Clinical information on pregnancies in this study is listed in Table 1. As expected from the PE pathogenesis, case pregnancies demonstrated lower gestational age, higher blood pressure and reduced fetal birth weight compared to controls (Table 1). PE was complicated by FGR in 61% of the cases.

3.2. The maternal HLA-G 14bp gene polymorphism 

All maternal blood samples (20 cases and 34 controls) were successfully genotyped for the HLA-G 14bp polymorphism. The overall frequency of the maternal HLA-G 14bp insertion allele irrespective of placental disease was 0.38, and is comparable to the frequency of 0.4 reported in Danish studies (Hviid et al., 2002, Hylenius et al., 2004). The maternal HLA-G 14bp genotype distribution in all cases and controls was 15% +14bp/+14bp, 39% −14bp/−14bp and 46% −14bp/+14bp, and showed expected Hardy–Weinberg equilibrium. Neither the maternal HLA-G 14bp allele frequency nor the HLA-G 14bp genotype distribution differed between cases and controls (Table 2). Similar negative findings were obtained when comparing primipara PE, severe PE or early-onset PE pregnancies to controls (Table 2 and data not shown).

Table 2. Maternal allele frequencies and genotype distribution of the HLA-G 14bp gene polymorphism in mothers with and without PE

		PEa		PE (primipara)b		Controls
		n	%	n	%	n	%
	Allele
	+14bp	15	37	10	36	26	38
	−14bp	25	63	18	64	42	62
	Total (n)	40		28		68
	Genotype
	+14bp/+14bp	2	10	2	14	6	18
	−14bp/−14bp	7	35	6	43	14	41
	−14bp/+14bp	11	55	6	43	14	41
	Total (n)	20		14		34

Case groups are compared to controls.

a Alleles: P=1.000; Fisher's exact test. Genotypes: P=0.596; Fisher's exact test. b Alleles: P=1.000; Fisher's exact test. Genotypes: P=1.000; Fisher's exact test.

3.3. The fetal HLA-G 14bp gene polymorphism and PE 

Genotyping of the fetal HLA-G 14bp polymorphism was successful in all but three umbilical cord blood samples, and 29 case and 28 control fetal genotypes were determined (Table 3). The overall frequency of the fetal HLA-G 14bp insertion allele was 0.44, and in accordance with Danish data (Hviid et al., 2002, Hylenius et al., 2004). The total fetal HLA-G 14bp genotype distribution in all cases and controls was 17% +14bp/+14bp, 30% −14bp/−14bp and 53% −14bp/+14bp, and showed expected Hardy–Weinberg equilibrium.

Table 3. Fetal allele frequencies and genotype distribution of the HLA-G 14bp gene polymorphism in pregnancies with and without PE

		PEa		PE (primipara)b		Controls
		n	%	n	%	n	%
	Allele
	+14bp	25	40	16	40	25	45
	−14bp	33	60	24	60	31	55
	Total (n)	58		40		56
	Genotype
	+14bp/+14bp	6	21	3	15	4	14
	−14bp/−14bp	10	34	7	35	7	25
	−14bp/+14bp	13	45	10	50	17	61
	Total (n)	29		20		28

Case groups are compared to controls.

a Alleles: P=1.000; Fisher's exact test. Genotypes: P=0.780; Fisher's exact test. b Alleles: P=0.681; Fisher's exact test. Genotypes: P=0.517; Fisher's exact test.

The fetal HLA-G 14bp allele frequency and genotype distribution did not differ between cases and controls (Table 3). Others have reported that the fetal HLA-G 14bp polymorphism is associated to primipara pre-eclamptic pregnancies (Hylenius et al., 2004, O’Brien et al., 2001), but no such association was observed in this study either comparing to all controls (Table 3) or to primipara controls only (data not shown). Similarly, no association was found between fetal HLA-G 14bp allele frequency or genotype distribution and subgroups of PE (severe or early-onset) (data not shown). The subgroup of primipara severe PE is reported associated with the fetal HLA-G 14bp genotype (Hylenius et al., 2004), and analysis of this specific group (n=16) gave one +14bp/+14bp, seven −14bp/−14bp and eight −14bp/+14bp fetal genotypes. In conclusion, the fetal HLA-G 14bp polymorphism did not appear to be related to PE in the Norwegian cohort.

3.4. The fetal HLA-G 14bp gene polymorphism and fetal growth 

Hviid (2004) have reported that the fetal HLA-G 14bp genotype is associated to birth weight. Birth weight normally increases with parity. The fraction of primiparas in this study varied among the fetal HLA-G 14bp genotypes: 33% of the fetal +14bp/+14bp genotypes, 55% of the fetal −14bp/−14bp genotypes and 53% of the fetal heterozygotes were from primipara pregnancies. The difference in parity was adjusted for and no association between fetal growth and fetal HLA-G 14bp genotype was observed (Table 4). Separate analysis of offspring from cases and controls suggested also that the fetal HLA-G 14bp genotype and fetal growth are not related (data not shown) and, furthermore, this appeared also to be the case when analysis was restricted to births at ≥38 weeks of gestation as performed in the study by Hviid (2004) data not shown).

Table 4. Fetal HLA-G 14bp genotype in relation to fetal growth

		Cases and controls
		n	BWDa
	Fetal genotype
	+14bp/+14bp	10	−0.92±1.07
	−14bp/−14bp	17	−1.15±1.59
	−14bp/+14bp	30	−0.77±1.86

Values are expressed as mean±S.D. Analysis by general linear model adjusted for parity. BWD, birth weight deviation; calculated as the difference between actual birth weight and expected birth weight (adjusted for gestational age and sex) and divided by the S.D. (to transform the variable birth weight to standard score).

a Adjusted P=0.684; General linear model, 2d.f.

3.5. Decidual gene expression of the HLA-G 14bp polymorphism 

The decidua basalis tissue of 30 cases and 29 controls were successfully analyzed by RT-PCR for gene expression from the HLA-G 14bp region. The 14bp insertion allele allows for two potential transcripts, the intact gene and an alternative spliced form lacking 92bp surrounding the 14bp insertion (Fujii et al., 1994, Hiby et al., 1999). The RT-PCR analysis revealed expression of one, two or three HLA-G specific transcripts (Fig. 1 and Table 5). The 226bp PCR product was derived from the HLA-G 14bp deletion allele, the 240bp fragment from the HLA-G 14bp insertion allele, and the short 148bp fragment identified the splice form of the HLA-G 14bp insertion allele (−92bp) (Fig. 1). All three potential HLA-G 14bp transcripts were detected with individual variations (Fig. 1 and Table 5), and simultaneous expression of all three transcripts was observed in 53% of the deciduas (Table 5). The two splice forms of the HLA-G 14bp insertion allele were always expressed together (Fig. 1A, lanes 4 and 7; B, lane 4; and Table 5).

View Large Image Download to PowerPoint

Fig. 1. Decidua basalis gene expression of the HLA-G 14bp gene polymorphism. RT-PCR with HLA-G specific primers showing HLA-G transcripts (marked 148, 226 and 240bp) from the gene region including the HLA-G 14bp gene polymorphism in decidua basalis tissue of cases (A) and controls (B). Data are representative of at least three independent RT-PCR experiments on all cases and controls. Corresponding fetal and maternal HLA-G 14bp genotypes are indicated. MW represents the molecular weight marker, and nd means no data obtained for specific genotypes.

Table 5. HLA-G 14bp gene expression in decidua basalis from pregnancies with and without PE

		PEa		Controls
		n	%	n	%
	mRNA transcripts
	+14bp/−92bp	5	17	4	14
	−14bp	11	37	8	28
	−14bp/+14bp/−92bp	14	47	17	59
	Total (n)	30		29

Cases are compared to controls.

a P=0.722; Fisher's exact test.

Decidual gene expression was compared in cases and controls and no difference in the HLA-G 14bp gene expression profile was detected (Table 5). Separate analysis of decidua basalis tissue from PE subgroups (primipara, severe or early-onset) did not reveal any significant associations to disease for the decidual HLA-G 14bp gene expression (data not shown).

3.6. HLA-G 14bp genotype in relation to HLA-G 14bp gene expression in decidua basalis 

Decidual HLA-G 14bp gene expression was compared to the corresponding maternal and fetal genotypes (exemplified in Fig. 1). For association to the fetal HLA-G 14bp genotype, 29 cases and 28 controls with both decidua basalis tissue and known fetal genotype were compared. Decidual HLA-G 14bp mRNA appeared to be linked directly to the fetal genotype (exemplified in Fig. 1). Of the 57 pregnancies compared, 56 deciduas (98.2%) showed an HLA-G 14bp mRNA profile in agreement with the corresponding fetal HLA-G 14bp genotype, and only one sample (1.8%) was in disagreement (Fig. 1A, last lane). For relation to the maternal HLA-G 14bp genotype, 19 cases and 20 controls were compared. Only 15 of the 39 deciduas (38.5%) showed an HLA-G 14bp gene expression correlating to the maternal HLA-G 14bp genotype and, in these individuals, maternal and fetal HLA-G 14bp genotypes were identical (exemplified in Fig. 1A, lanes 3–5; B, lane 4). The single incongruent decidual mRNA (only the −14bp transcript) did not match either corresponding fetal genotype (+14bp/+14bp) or maternal genotype (−14bp/+14bp) (Fig. 1A, last lane) and, in lack of obvious biologic explanations, we ascribe this single result to probable experimental error. In all other individuals, the fetal −14bp/−14bp genotype was associated always with one HLA-G 14bp transcript in decidua basalis and the fetal +14bp/+14bp genotype was associated with two transcripts. The heterozygous fetal HLA-G 14bp genotype (−14bp/+14bp) was, without exception, associated with the expression of all three HLA-G 14bp mRNA transcripts (exemplified in Fig. 1).