The genetics of feto-placental development: A study of acid phosphatase locus 1 and adenosine deaminase polymorphisms in a consecutive series of newborn infants

Background Acid phosphatase locus 1 and adenosine deaminase locus 1 polymorphisms show cooperative effects on glucose metabolism and immunological functions. The recent observation of cooperation between the two systems on susceptibility to repeated spontaneous miscarriage prompted us to search for possible interactional effects between these genes and the correlation between birth weight and placental weight. Deviation from a balanced development of the feto-placental unit has been found to be associated with perinatal morbidity and mortality and with cardiovascular diseases in adulthood. Methods We examined 400 consecutive newborns from the Caucasian population of Rome. Birth weight, placental weight, and gestational length were registered. Acid phosphatase locus 1 and adenosine deaminase locus 1 phenotypes were determined by starch gel electrophoresis and correlation analysis was performed by SPSS programs. Informed verbal consent to participate in the study was obtained from the mothers. Results Highly significant differences in birth weight-placental weight correlations were observed among acid phosphatase locus 1 phenotypes (p = 0.005). The correlation between birth weight and placental weight was markedly elevated in subjects carrying acid phosphatase locus 1 phenotypes with medium-low F isoform concentration (A, CA and CB phenotypes) compared to those carrying acid phosphatase locus 1 phenotypes with medium-high F isoform concentration (BA and B phenotypes) (p = 0.002). Environmental and developmental variables were found to exert a significant effect on birth weight-placental weight correlation in subjects with medium-high F isoform concentrations, but only a marginal effect was observed in those with medium-low F isoform concentrations. The correlation between birth weight and placental weight is higher among carriers of the adenosine deaminase locus 1 allele*2, which is associated with low activity, than in homozygous adenosine deaminase locus 1 phenotype 1 carriers (p = 0.04). The two systems show a cooperative effect on the correlation between birth weight and placental weight: the highest value is observed in newborns carrying adenosine deaminase locus 1 allele*2 and acid phosphatase locus 1 phenotypes with medium-low F isoform concentration (p = 0.005). Conclusion These data suggest that zygotes with low adenosine deaminase locus 1 activity and low F activity may experience the most favourable intrauterine conditions for a balanced development of the feto-placental unit.


Background
We have recently described a cooperative interaction between ACP 1 (acid phosphatase locus 1) and ADA 1 (adenosine deaminase locus 1) genetic polymorphisms concerning their effects on the susceptibility to spontaneous primary repeated miscarriages: women carrying the ADA 1 *2 and ACP 1 *C alleles show the lowest susceptibility to repeated miscarriages [1]. Both systems share important effects on glucose metabolism and immunological function. These observations prompted us to search for a possible cooperative interaction between the two systems regarding their effects on developmental parameters during intrauterine life.
It is likely that a balanced growth of the two portions of the feto-placental unit (i.e. without the prevalence of placental on the fetal part or vice versa) represents an advantage for fetal development. The birth weight/placental weight ratio (BW/PW) has been found to be correlated with perinatal morbidity and mortality and with cardiovascular disease in adulthood [2,3]. There is evidence that in addition to maternal factors and socioeconomic status, genetic factors also influence the ratio BW/PW [3][4][5]. In a recent note we have proposed the correlation between BW and PW as an index of balanced development of the fetoplacental unit and have shown that this correlation is influenced by ACP 1 phenotype [6].
The ACP 1 genetic polymorphism ACP 1 also called low molecular weight phospho-protein tyrosine phosphatase (LMPTP) is an enzyme controlled by a locus on chromosome 2 showing three common alleles: ACP 1 *A, ACP 1 *B, ACP 1 *C. These three alleles are associated with different enzymatic activities [7]. Activity of ACP 1 phenotypes are in te following order: A < BA < B = CA < CB < C [8].
Each allele at the ACP 1 locus encodes two isoforms, called F (fast) and S (slow) [9,10]. ACP 1 B and BA show a medium-high F isoform activity while CB, A, CA and C phenotypes show a medium-low F isoform activity. ACP 1 C, CA and CB show a much higher activity of S isoform as compared to other ACP 1 phenotypes [7,10].
Two important functions have been suggested for ACP 1 : flavin-mono-nucleotide phosphatase activity and tyrosine phosphatase activity [11][12][13]. Catalysing the conversion of flavin-mononucleotide (FMN) to riboflavin, ACP 1 may have a role in regulating the cellular concentration of flavin-adenine-dinucleotide (FAD), flavo-enzyme activity and energy metabolism. As a phosphotyrosine phosphatase, the enzyme may have an important role in cellular growth regulation and in modulation of glycolytic rate through the control of receptor activities and of band 3 protein phosphorylation status [ [12,14] and [15]].
Recently it has been shown that ACP 1 specifically dephosphorylates the negative regulatory Tyr-292 of ZAP-70, thereby counteracting inactivation of ZAP-70. The ZAP-70 protein-tyrosine kinase plays a central role in signalling from the T cell receptor. Thus, these results indicate that ACP 1 strengthens T cell receptor signalling [16].
The ADA 1 genetic polymorphism ADA 1 is a polymorphic enzyme present in all mammalian tissues [17]. It is controlled by a locus with two codominant alleles ADA*1 and ADA*2 located on the long arm of chromosome 20. The corresponding three common ADA 1 phenotypes have different enzymatic activities: the ADA 1 1 phenotype is 15% more active than the ADA 1 2/1 phenotype and 30% more active than the ADA 1 2 phenotype, which is very rare [18].
ADA 1 catalyses the irreversible deamination of adenosine to inosine. Red Blood Cells (RBC) are in equilibrium with freely diffusing adenosine [19], pointing to an important role for this enzyme in the regulation of adenosine concentration.
Current interest has been focused on a wide variety of effects produced by adenosine via activation of cell surface adenosine receptors [20,21]. Adenosine counteracts insulin action in the liver by activating A2B receptors [22]. Adenosine seems to facilitate insulin action in adipocytes.
The adenosine deaminase complex protein [23] (ADPC) is identical with CD26, a T cell activating antigen and with a glycoprotein present in epithelial cells of various tissues. Recent data suggest that ADA 1 and CD26 are co-localized on the T cell surface but not inside cells.
Cells expressing ADA 1 and CD26 on the surface are much more resistant to the inhibitory effects of adenosine. These data suggest that ADA 1 on the cell surface is involved in an important immunoregulatory mechanism by which released ADA 1 binds to the cell surface of CD26, and this complex is capable of reducing the local concentration of adenosine [24].
In the present paper we have performed a more detailed analysis of the effect of ACP 1 polymorphism and have extended this study to the ADA 1 polymorphism. On the basis of the observations on women with repeated spontaneous miscarriages we would expect an optimal developmental context in zygotes carrying the ADA 1 *2 and ACP 1 *C alleles.

Methods
In the present study we examined 400 consecutive newborn infants from healthy puerperae. All infants were Caucasian from the population of Rome. Birth weight and placental weight (wet, untrimmed) were registered in the delivery room. Gestational length was estimated from the date of the last menstrual period and checked against Dubowitz score as an additional index of neonatal maturity. Multiples were excluded. Mode of delivery does not alter the phenotype of enzyme considered.
The data presented in the paper were collected a few years ago and at that time there was not an established Ethical Committee. The project was discussed and approved in the Department. Informed verbal consent to participate in the study was obtained from the mothers. This has been recently (April 28, 2008) approved by the Institutional Ethical Committee.
Newborn blood samples were obtained from the placental side of the umbilical vein after umbilical cord section. The ACP 1 phenotype was determined in 361 newborns by starch gel electrophoresis on red blood cell hemolysates according to Harris and Hopkinson [25]. The acid phosphatase pattern is revealed by a solution of phenolphtalein diphosphate: the addition of ammonium solution reveals the area where phenolphtalein has been liberated in the areas of gel where ACP 1 activity is present. In European populations the presence of three common alleles *A, *B and *C determines the occurrence of six phenotypes: A, AB, B, AC, BC and C. Each of the homozygous A, B and C phenotypes are composed of two fractions, F and S, corresponding to fast and slow components of the electrophoretic pattern. Heterozygous phenotypes have a pattern corresponding to a mixture of homozygous types.
The ADA 1 phenotype was determined by starch gel electrophoresis on red blood cell hemolysates according to Spencer et al [26]. Inosine produced at the sites of ADA 1 activity is converted in hypoxanthine in the presence of nucleoside phosphorylase and phosphate. The hypoxanthine is then oxidized by the action of xanthine oxidase, and during this reaction the tetrazolium salt MTT is reduced in the presence of phenazine methosulphate to a blue insoluble formazan. In the ADA 1 1 type there are three regularly spaced components which exhibit decreasing staining intensity in order of their anodal electrophoretic mobilities. In the ADA 1 2 type there are also three isozymes and their relative intensities and relative electrophoretic mobilities are very similar to those of the ADA 1 1 pattern. The difference between ADA 1 1 and ADA 1 2 is that the ADA 1 2 pattern is appreciably slower than the ADA 1 1 pattern. The pattern exhibiting four isozymes, designated ADA 1 2/1, has the appearance of a mixture of ADA 1 1 and ADA 1 2 patterns.
In the last few years, in our laboratory, determination of ADA 1 and ACP 1 genotypes has been performed routinely on DNA. In our laboratory the comparison of classical with DNA methods has shown practically no differences between phenotypic and genotypic classifications. On a sample of 50 subjects in which ACP 1 and ADA 1 phenotypes were determined by DNA and classical methods only one difference was observed for ACP 1 and no difference for ADA 1 .
Correlation analysis was performed by SPSS programs. Differences between correlation coefficients were evaluated according to Snedecor and Cochran [27]. The distribution of ACP 1 phenotypes among newborns does not differ statistically from Hardy-Weinberg expectation. Table 1 shows demographic parameters of the sample study. Table 2 shows the distribution of ACP 1 phenotypes and developmental parameters for each phenotype. No statistical significant difference among ACP 1 phenotypes is observed for BW, PW and gestational duration. CA phenotype shows low values for all parameters, but these values are not statistically different from those of other ACP 1 phenotypes. BW-PW correlation analysis shows highly significant differences among ACP 1 phenotypes. The highest correlation coefficient is observed for CA phenotype and the lowest for B phenotype. Figures 1, 2, 3, and 4 illustrate the relationship between BW-PW correlation and relevant ACP 1 parameters: F and S activity, F/S activity ratio and total activity. The BW-PW correlation is negatively associated with F concentration (Fig 1) and F/S ratio (Fig 3). A, CA and CB phenotypes that share a medium-low F activity and F/S ratio have a high BW-PW correlation, while the B phenotype, which has the highest F activity and F/S ratio, has a low BW-PW correlation. No association is observed for S isoform concentration (Fig 2) and ACP 1 total activity (Fig 4). Table 3 shows the effect of smoking, maternal age, gestational age, parity and gender on the relationship between ACP1 and BW-PW correlation. For this analysis two classes of ACP1 phenotypes have been considered: A, CA, CB phenotypes with medium-low F activity, and B and BA phenotypes with medium-high F activity. In A, CA, CB  Table 4 shows the distribution of ADA 1 phenotypes and developmental parameters for the ADA 1 1 phenotype and for carriers of the ADA 1 *2 allele. The BW-PW correlation coefficient is higher among carriers of the ADA 1 *2 allele than among ADA 1 1 phenotype carriers (p = 0.04). No significant difference between ADA 1 phenotypes has been observed for BW, PW and gestational duration. Table 5 shows the effect of smoking, maternal age, gestational age, parity and gender on the relationship between ADA 1 and BW-PW correlation. Significant effects of these variables were observed in the correlations of BW-PW in both ADA 1 subjects and in carriers of the ADA 1 *2 allele.

Results
The BW-PW correlation is influenced by environmental and developmental variables, but as shown in tables 3 and 5 the differential effects of ACP 1 and ADA 1 phenotypes on this correlation do not appear in general to be significantly influenced by these variables. Table 6 shows the correlation between BW and PW according to the joint ACP 1 -ADA 1 phenotype. Both ADA 1 and ACP 1 phenotypes were determined in 327 infants. There is a highly significant difference among joint phenotypes (p = 0.000): the highest correlation is observed in subjects who carry A or CA or CB phenotypes and the ADA 1 *2 allele, while the lowest correlation is observed when the ADA 1 1 phenotype is associated with B or BA phenotypes.

Discussion and conclusion
The present data suggest that foetuses with low ADA 1 activity, associated with medium-low ACP 1 F isoform activity, have the best correlation between BW and PW, suggesting a most favourable situation for the development of the feto-placental unit. These observations agree with those expected on the basis of previous data on women with repeated spontaneous miscarriages that demonstrated a cooperative protective effect of ADA 1 *2 and ACP 1 *C alleles against fetal loss. Thus, the two lines of evidence support the hypothesis that foetuses with low ADA 1 activity and low ACP 1 F isoform activity have a balanced development of feto-placental unit and a higher probability of survival compared to other foetuses. Significance of differences among means (*) refers to Variance analysis. Significance of differences among correlation coefficients (**) has been calculated according to Snedecor and Cochran. The relationship between F/S isoform concentration and BW-PW correlation Figure 3 The relationship between F/S isoform concentration and BW-PW correlation. The term BW-PW correlation expresses the correlation between birth weight and placental weight. ACP1 is the acid phosphatase locus 1. A, B, C, BA, CA, CB are the ACP 1 phenotypes. In abscissa F/S isoform concentrations of each ACP 1 phenotype are also reported. The rank correlation coefficient according to Spearman (27) between BW-PW correlation and F/S isoform concentration is r s = -0.9, p < 0.05. The relationship between F isoform concentration and BW-PW correlation Figure 1 The relationship between F isoform concentration and BW-PW correlation. The term BW-PW correlation expresses the correlation between birth weight and placental weight. ACP1 is the acid phosphatase locus 1. A, B, C, BA, CA, CB are the ACP 1 phenotypes. In abscissa F isoform concentrations of each ACP 1 phenotype are also reported. The rank correlation coefficient according to Spearman (27) between BW-PW correlation and F isoform concentration is r s = -1, p < 0.01.

BW-PW correlation
The relationship between S isoform concentration and BW-PW correlation Figure 2 The relationship between S isoform concentration and BW-PW correlation. The term BW-PW correlation expresses the correlation between birth weight and placental weight. ACP1 is the acid phosphatase locus 1. A, B, C, BA, CA, CB are the ACP 1 phenotypes. In abscissa S isoform concentrations of each ACP 1 phenotype are also reported. The rank correlation coefficient according to Spearman (27) between BW-PW correlation and S isoform concentration is r s = 0.3, p not significant.

BW-PW correlation
The relationship between F+S isoform concentration and BW-PW correlation Figure 4 The relationship between F+S isoform concentration and BW-PW correlation. The term BW-PW correlation expresses the correlation between birth weight and placental weight. ACP1 is the acid phosphatase locus 1. A, B, C, BA, CA, CB are the ACP 1 phenotypes. In abscissa F+S isoform concentrations of each ACP 1 phenotype are also reported. The rank correlation coefficient according to Spearman (27) between BW-PW correlation and F+S isoform concentration is r s = 0.3, p not significant.

BW-PW correlation
The exact mechanism underlying the statistical association of ADA 1 and ACP 1 with the BW-PW correlation is not known at present. An immunological mechanism is supported by the well known relationship between ADA 1 and immune diseases and between ACP 1 and T cell activation. A relative depression of T cell activation due to low level of F ACP 1 isoform and to higher concentration of adenosine (due to the low activity of ADA 1 *2 carriers) could modulate the feto-maternal immunological relationship resulting in a balanced development of the two portions of feto-placental unit.
A metabolic mechanism may be operative in which ACP 1 , acting as phosphotyrosine phosphatase, could have an important role in the modulation of glycolytic rate through the control of insulin receptor activity and of band 3 protein phosphorylation status. Additionally, catalysing the conversion of flavin-mononucleotide (FMN) in riboflavin, the enzyme may influence flavo-enzyme activity and energy metabolism [7]. In turn, with respect to ADA 1 activity, recent studies have shown that adenosine counteracts insulin action in the liver by activating A2B receptors [20][21][22]. On the basis of these actions on glucose metabolism, ACP 1 F isoform activity coupled with low ADA 1 activity could have favourable effects on the development of the feto-placental unit. Significance of difference between correlation coefficients has been calculated according to Snedecor and Cochran. p* refers to significance of correlation coefficient. p** refers to significance of difference between the two classes. For "All subjects" the difference of correlation coefficients between (B+BA) vs (A+CA+CB) i.e. 0.209 vs 0.687 is highly significant: p = 0.002.
Regarding the effect of genetic variability of ACP 1 on the correlation of BW with PW, it is interesting to speculate on the possible selective advantage of the *A and *C alleles over the *B allele. The *B allele is the most frequent in all human populations, and the *A allele is present with variable frequencies in all major ethnic groups, while *C allele is present with appreciable frequencies only in Caucasians.
Our data suggest that the optimal BW-PW correlation is seen in carriers of the ACP 1 *C allele and in the homozygous A phenotype (table 4 and Fig 1), while the heterozygous BA phenotype shows an intermediate value between B and A (Fig 1). Interestingly, in A, CA and CB phenotypes the BW-PW correlation is hardly influenced by environmental circumstances, while in B and BA phenotypes the environmental variables exert considerable effects on this correlation. Thus, ACP 1 *A and *C variants could have a selective advantage during intrauterine life on the fundamental ACP 1 *B allele. This might have contributed to an increase in frequencies of the ACP 1 *A and *C alleles to polymorphic values and could presently contribute to maintenance of the ACP 1 polymorphism in human populations.

Competing interests
The authors declare that they have no competing interests. Correlation between birth weight and placental weight (r) 0.289 0.552 0.040** Significance of r (p) 0.000 0.000 Significance of difference between means (*) refers to Variance analysis. Significance of difference between correlation coefficients (**) has been calculated according to Snedecor and Cochran