Plasminogen activator inhibitor type 1 (PAI1) and platelet glycoprotein IIIa (PGIIIa) polymorphisms in Black South Africans with pre-eclampsia

Abstract

Background. Association of fibrin abnormalities with pre-eclampsia prompted this study to examine whether polymorphisms in the plasminogen activator inhibitor Type 1 and platelet glycoprotein IIIa genes constitute risk factors for this condition.

Methods. A group of 151 Black Zulu-speaking pre-eclamptics was examined for 4G/5G plasminogen activator inhibitor Type 1 and Pl^A1/A2 platelet glycoprotein IIIa polymorphic alleles using standard techniques. Results were compared with those found in 217 ethnically matched healthy normotensive pregnant women who had normal full-term gestations.

Results. Pre-eclamptic patients had a slightly higher frequency of the 4G plasminogen activator inhibitor Type 1 allele (15%) compared with the controls (12%); this was reflected also in the heterozygote frequency (28% and 22%) for the patients and the controls, respectively. These differences were not significant. Only 2% of this population was found to be homozygous for the 4G allele. No differences were observed in the platelet glycoprotein IIIa polymorphism genotype and allele frequency distribution between the patients and the controls.

Conclusions. Neither the 4G allele of the plasminogen activator inhibitor Type 1 nor the Pl^A2 allele of the platelet glycoprotein IIIa have any significant role as risk factors in the patho-etiology of pre-eclampsia in Black South Africans, although these genes cannot yet be excluded as contributory to this disorder. It is possible that the underlying causes of pre-eclampsia may vary between different ethnic populations.

Abbreviations:

PE

pre-eclampsia

PAI1

plasminogen activator inhibitor Type 1

PGIIIa

platelet glycoprotein IIIa

MTHFR

methylenetetra- hydrofolate reductase.

Pre-eclampsia (PE) is commonly associated with fibrin abnormalities and occlusive lesions within the intervillous or spiral artery of the placenta (1). Several factors are known to contribute to the incidence of thrombotic disorders, in particular defects in the coagulation pathway, and a number of specific genetic aberrations are reported to be associated with vascular disease in general. These include the Factor V Leiden, prothrombin 20210 G (→)A, thrombomodulin A455V and methylenetetrahydrofolate reductase (MTHFR) 677 C (→)T mutations (2–5). All these variants, and several others, have also been reported to predispose towards PE (6–9), although these claims are not unchallenged (10,11).

Recently the plasminogen activator inhibitor Type 1 (PAI 1) gene has also been implicated in vascular/thrombotic events. There have been reports of elevated levels of PAI 1 in the plasma and placenta of pre-eclamptic women (12), as well as significantly increased expression of PAI 1 mRNA (13). Increased plasma levels of PAI 1 were subsequently shown to be related to a single base pair guanine deletion/insertion (4G/5G) polymorphism in the promoter region of the PAI 1 gene located − 675 bp from the transcription start site. The allele with the single G insertion (5G) contains an additional repressor protein-binding site, in the absence of which (4G) PAI 1 transcription is increased, leading to hypofibrinolysis and thrombophilia, and a significant association with myocardial infarction (14,15). The relationship of 4G/4G homozygous status to pre-eclampsia has subsequently been assessed in two studies (16,17), with both groups reporting the homozygous 4G/4G genotype to be one of the risk factors for PE in both Caucasian and Japanese populations. These observations have not yet been confirmed or refuted, nor have any studies been carried out on other ethnic groups.

In the current study we analyzed the PAI 1 hypofibrinolytic 4G/5G mutation in a group of Zulu-speaking Black South African women with PE to assess its impact on this disorder. Following a report that the concurrent carrier status of the PAI 1 gene 4G allele with the Pl^A2 allele of the platelet glycoprotein IIIa (GPIIIa) gene conferred an increased risk for coronary artery disease (18), it was decided to examine the GPIIIa polymorphic allele as well. Mutations in other genes purported to be associated with PE, such as the Factor V Leiden, prothrombin (20210 G (→)A), thrombomodulin (A455V) and MTHFR (677 (→)T), have been reported previously in the same cohort (19; Hira et al., personal communication).

Materials and methods

One hundred and fifty-one Black Zulu-speaking South African women with documented severe PE were recruited from our labor ward high-care complex where all high-risk patients are monitored. Severe pre-eclampsia was defined by a BP > 160/110 mmHg (Korokoff 5) on two separate occasions 6 h apart, +++ proteinuria on dipstix, and a platelet count > 150 × 10⁹/L. Antenatal and delivery history were obtained from patient records. All women with PE were normotensive and aproteinuric within 2 weeks of delivery. The control group comprised 217 age-matched women of the same population who were recruited from the postdelivery wards after review of their records to ensure the absence of any antenatal, intrapartum, or postpartum complications. Women with known endocrine or metabolic disorders or a history of hypertension in a previous pregnancy were excluded from this group. Only indigenous Zulu-speaking Black African women were included in the study and those with known nonindigenous relatives were excluded. Institutional ethical approval was obtained from the Ethics Committee; all participation was based on informed voluntary consent before blood samples were obtained.

DNA analysis

DNA was isolated by standard techniques. Detection of the PAI 1–675 4G/5G polymorphism was performed under standard PCR conditions with the primers 5′- CCAGCACACC TCCAACCT-3′ (F) and 5′-CCTCATCCCTGCCATGTG-3′ (R), and 1.5 mmol/L of MgCl₂ using a fluorescent dUTP (Applied Biosystems, Foster City, California, USA) at an annealing temperature of 53 °C. After purification using glass fiber filters (Roche High PCR Product Purification kit, Roche, Mannheim, Germany), the PCR products were analyzed on an ABI Prism 310 Genetic Analyzer using Genescan Software (v. 3.1.2) (Applied Biosystems, Foster City, California, USA) to determine the exact fragment lengths. Detection of the T→C transition at position 196 in exon 2 of the GPIIIa gene, which creates a Msp I cutting site, was based on the method of Ridker et al. (20). The primers used were 5′-TGCTGGACTT CTCTTTGGGC-3′ (F) and 5′-TCAGGTCTC TCCCCGCAAAG-3′ (R), in standard PCR with 1.5 mmol/L of MgCl₂ at an annealing temperature of 58 °C. The alleles were designated Pl^A1 (T) and Pl^A2 (C). Heterozygote DNA verified by nucleic acid sequencing was used as a control for each of these tests in all PCR runs.

Statistics

The allele frequencies were estimated by gene counting. The χ²-test was used to examine allele and genotype differences within the patient and control groups; the variance was expressed in terms of 95% confidence intervals. In all groups the distribution of alleles was tested for the Hardy–Weinberg equilibrium. For reference purposes, we have also included the results of PAI 1 4G/5G and GPIIIa Pl^A1/A2 polymorphic analyses on a cohort of healthy Caucasian blood donors performed by our laboratory.

Results

The PAI 1 and GPIIIa gene polymorphism distributions are shown in Table I. The genotype frequencies for both polymorphisms were compatible with the Hardy–Weinberg equilibrium. Pre-eclamptic patients had a slightly higher frequency of the PAI 1 4G allele than the controls, as reflected in the higher heterozygote rate (28% vs. 22%). However, the overall differences between the two groups, with respect to both genotype and allele frequencies, were not statistically significant. Of note, is the very low (2%) incidence of 4G/4G homozygotes seen in this population compared with 19% seen in Caucasians (Table I). No differences were observed in the GPIIIa Pl^A1 and Pl^A2 genotype and allelotype distribution between the patients and the controls. Furthermore, there was no concurrent relationship status between the two genetic variants examined, and no individuals were homozygous for both alleles.

Table I. PAI 1 and GPIIIa genotypes and allele frequencies in Black patients with pre-eclampsia and control subjects

Polymorphism	Study group (n = 151)	Control group (n = 217)	χ^2*	p	OR (95% CI)	Caucasian group
PAI 1
Genotype						(n = 150)
4G/4G	2 (2%)	2 (2%)	0.02	0.885	1.4 (0.1–20.1)	29 (19%)
4G/5G	42 (28%)	48 (22%)	1.27	0.260	1.4 (0.8–2.3)	67 (45%)
5G/5G	107 (70%)	167 (76%)	1.43	0.231	0.7 (0.4–1.2)	54 (36%)
Allele
4G	46 (15%)	52 (12%)	1.36	0.243	1.3 (0.8–2.1)	294 (49%)
5G	256 (85%)	382 (88%)	–	–	0.8 (0.5–1.2)	310 (51%)
GPIIIa
Genotype						(n = 207)
Pl^A1/A1	119 (79%)	160 (74%)	0.99	0.320	1.3 (0.8–2.3)	151 (73%)
Pl^A1/A2	30 (20%)	52 (24%)	0.64	0.423	0.8 (0.5–1.3)	54 (26%)
Pl^A2/A2	2 (1%)	4 (2%)	0.00	0.975	0.7 (0.06–5.1)	2 (1%)
Allele
Pl^A1	268 (89%)	372 (86%)	1.8	0.276	1.3 (0.8–2.1)	356 (86%)
Pl^A2	34 (11%)	62 (14%)	–	–	0.8 (0.5–1.2)	58 (14%)

• * All χ² comparisons had 1 degree of freedom.

Discussion

The main finding of this study was that in a population of Black South African women the frequency of neither the hypofibrinolytic 4G allele of the PAI 1 gene nor the Pl^A2 variant of the GPIIIa gene differed between subjects with clinically diagnosed PE and a group of age and ethnically matched controls who had full-term normotensive gestations. This suggests that in this ethnic population at least, neither of these two polymorphisms serves as risk factors for PE. These findings are at some variance with two other reports in which the PAI 1 4G/4G genotype was shown to significantly increase the risk of PE in both Caucasians and Japanese (16,17). In the Caucasian study the patient group comprised a combined population of complicated pregnancies (n = 94), but this included only 31 patients with severe PE, five of whom had eclampsia and five who had hemolysis, elevated liver enzymes, and low platelets syndrome. No information on the findings within these groupings was given, most likely because the numbers were small. The Japanese study comprised 115 pre-eclamptic patients.

There are, to the best of our knowledge at this point, no other data available on the relationship between PAI 1 4G/4G homozygosity and the risk of PE. Neither do there appear to be any studies examining the GPIIIa Pl^A2 allele in PE. Our study has attempted to address this in a large (n = 151) ethnically discrete Black Zulu-speaking South African population in which PE is the most common cause of maternal and fetal morbidity and mortality, affecting 18% of pregnancies (21). Control subjects (n = 217), all of whom were sampled postdelivery after normal pregnancies, were drawn from the same ethnic pool and did not differ with respect to either age or parity.

Of immediate note in this study was the markedly lower frequency of the PAI 1 4G allele (12–15%) in this African population compared with 30% in the Japanese (22) and approximately 50% reported in Caucasians by Hooper et al. (23) and ourselves in the present study (Table I). Given this low frequency, coupled with the 2% incidence of homozygosity, it would seem unlikely that the hypofibrinolytic effect of the PAI 1 4G allele has any major impact on Black African women, at least in the patho-etiolgy of PE, a condition that affects approximately 18% of pregnancies (21). This does not however, exclude the possibility that the PAI 1 gene may be involved through some other mechanism, particularly as high levels of plasma and placental PAI 1, as well as increased PAI 1 expression, are strongly associated with PE (12). Other polymorphisms in the PAI 1 gene have been reported, such as a A→G substitution at − 844 in the consensus sequence-binding site (24,25), and need to be investigated further in relation to PE.

In the study by Glueck et al. (16), individuals homozygous for the PAI 1 4G allele, who also carried the Factor V Leiden mutation, prothrombin (20210 G→A) or were homozygous for the MTHFR 677C→T mutation, were reported to have a significantly higher risk of pregnancy complications. There have been similar findings in venous thrombosis in which the −844 AA genotype was found to be a risk factor in Factor V Leiden carriers (26). These observations suggest that defects in two or more separate genes can contribute additively. In the Black population studied here however, both the Leiden and prothrombin mutations are absent (Hira et al. 2002, personal communication), while MTHFR 677C→T homozygosity is rare (19), implying involvement of other mutations in these genes, or possibly even entirely different genes.

With respect to the GPIIIa gene, there does not seem to be a role for the Pl^A2 allele as a predisposing factor in PE. The frequency of this allele in this population was similar to that reported in Caucasians both by others (20) and in this study, and no differences were observed between our patient and control groups. Despite the report by Pastinen et al. (18), in which the concurrent carrier status of the GPIIIa Pl^A2 and PAI 1 4G alleles conferred a high risk for the development of myocardial infarction in a Finnish population, suggesting additive risk, no similar observation for PE emerged in our study.

Results demonstrate that neither the PAI 1 4G homozygous genotype nor the GPIIIa Pl^A2 allele play a contributory role on PE. The etiology of this disorder is undoubtedly complex. It may vary between ethnic populations, but seems likely to result from multigenetic defects that interplay and may be moderated by lifestyle and environmental factors.

Address for correspondence:
R. J. Pegoraro
Department of Chemical Pathology
Nelson R Mandela School of Medicine
University of Natal
Private Bag 7
Congella, 4013
South Africa
e-mail: [email protected]