Main

In humans, the perinatal period is characterized by several switches in Hb composition. The major switch in Hb type occurs late during fetal development, when fetal Hb (α2γ2), the most important Hb of intrauterine life is gradually replaced by adult Hb (γ2β2)(1). Understanding the genetic and epigenetic mechanisms of fetal to adult Hb switch has fundamental as well as therapeutic implications. The changes in γ- and β-globin gene activity have been assessed by quantifying their mRNA levels, and we have shown that they correlate very precisely with the amounts of γ- and β-globins being synthesized. This recently published finding demonstrated that the quantification of γ-globin mRNA as a percent of a non-α-globin mRNA closely reflects globin chain synthesis and can be considered an accurate measure of HbF synthesis during physiologic conditions(2).

The γ-globin expression has been shown to be greater than the normal range in infants during the perinatal period under certain pathologic conditions, such as placental insufficiency(3), maternal diabetes(4), severe anemia(5), BPD(6) and CCHD(7). Erythropoietic stress(8,9), which describes states of red blood cell expansion that occurs in response to proliferative signals, is the explanation of this phenomenon. Hypoxemia, by triggering an expansion of erythropoiesis, has been shown to increase the production of HbF(7,10,11).

Because of documentation that there exists an accurate quantitative relationship between globin mRNA levels and globin peptide synthesis under normal physiologic conditions(2), we considered it important to determine whether this quantitative relationship is maintained when a reactivation of HbF synthesis occurs during pathologic conditions that can stimulate stress erythropoiesis. Therefore, a study was planned to examine the quantitative relationship between the mRNAs encoding globins and their translation into proteins during conditions that are known to cause higher than normal levels of HbF synthesis, such as in newborn infants of diabetic mothers and in newborn IUGR infants of toxemic mothers. For similar reasons, blood samples from infants with BPD and CCHD were also analyzed.

The data obtained from this study would determine whether mRNA globin levels can give an accurate reflection of the percent of HbF synthesis during conditions of stress erythropoiesis. If an accurate quantitative relationship between the globin mRNAs and the globin peptide synthesis during pathologic conditions were confirmed, the mRNA quantification, which requires a minimal amount of blood, could then be used as a rapid method for measuring levels of HbF synthesis. This information would become very valuable should it be necessary to determine the levels of HbF being produced for clinical or diagnostic purposes.

METHODS

Study population. To determine the relationship between globin mRNAs and their protein synthesis in infants expected to have increased levels of HbF for their age of development, 15 infants were included in the study. Cord blood samples were obtained from five newborns from insulin-dependent diabetic mothers and two IUGR newborn infants from toxemic mothers. Blood samples were also obtained from seven infants with CCHD (2-9 y of age) and three infants with severe BPD (born at 27.5, 29, and 30 wk of gestation and sampled at 36.5, 43.5, and 48.5 wk postconceptional age). Also included were seven normal adult controls. All procedures were approved by the Institutional Review Board of St. Justine Hospital.

Measurement of de novo globin synthesis and mRNAs. For the quantification of the de novo protein synthesis and their globin mRNAs, recently published methodology was used(2). In summary, the synthesis of α-, β-, γ-globins in reticulocytes was determined by the incorporation of [3H]-leucine followed by the separation and quantification of the polypeptides on a C4-reverse phase HPLC, whereas the relative proportions of the mRNAs of α-, β-, γ-globins were measured by RNase protection assay. Total RNA isolated from blood (50 µL) was incubated with [32P]-labeled cRNA probes of α-, β-, γ-globins, followed by RNase A/T1 digestion, separation of the protected fragments, and densitometry of the radioactive bands by phosphorimaging(2).

Statistical analysis. Data are expressed as means ± SD. An unpaired t test was used to compare the means of two groups; p < 0.05 was considered statistically significant. Linear regression analysis was used to determine relationships between variables.

RESULTS

Figure 1 demonstrates the levels of HbF synthesis based on γ/(γ + β)-globin synthesis from the IUGR, IDM, and BPD infants included in this study, in relation to postconceptional age. The regression line and 95% confidence interval limits that are shown in this figure are from the HbF synthesis of normal infants, previously published(12). All but two of the infants included in this figure had HbF synthesis levels greater than the 95th percentile. The BPD infant who had a postconceptional age of 48.5 wk had a level of HbF synthesis (57.1%) that was more than 2 SD above the values previously reported(6).

Figure 1
figure 1

A comparison of [γ/(γ + β)] from erythropoietic stressed infants (IUGR, BPD, and IDM) correlated with their postconceptional age. The regression line (–) and 95% confidence interval limits (–––) are from the HbF synthesis of normal infants from a previous study(12).

Full size image

Figure 2 demonstrates the levels of [γ/(γ + β)]-globin mRNAs obtained in CCHD infants, compared with controls. The data illustrate a significant difference between the two groups. Four of the seven infants with CCHD had amounts of [γ/(γ+β)]-globin mRNAs >2 SD above those of controls.

Figure 2
figure 2

A comparison of the [γ/(γ + β)]-globin mRNAin infants with CCHD compared with controls. Data are expressed as mean ± SD. * denotes a significant difference, p = 0.04.

Full size image

A comparison of the proportions of de novo globin chain synthesis and the mRNAs encoding these globins from all 17 infants included in the study is shown in Figure 3. There was a highly significant correlation (r2 = 0.994; p < 0.001). The regression equation was y = 0.94x + 4.45, and the SEM of the estimate was 2.8%. When only the 12 infants who had increased HbF synthesis are included in the above comparison (four CCHD infants whose levels were >2 SD of the mean of the controls and eight infants whose levels were >95% confidence limits of normal newborn infants), the correlation remains unchanged (r2 = 0.995, p < 0.001).

Figure 3
figure 3

Correlation between globin synthesis and the mRNAs encoding the globins from infants during stress erythropoiesis: r2=0.994, p < 0.001.

Full size image

DISCUSSION

Elevated levels of fetal Hb are observed in humans in two significantly different contexts. The first is in the setting of a diverse group of genetic mutations that are unified phenotypically by persistent synthesis of HbF after birth, known collectively as hereditary persistence of HbF(13). The second involves the stimulation of the quiescent fetal genes in erythroid progenitors, as seen in stress erythropoiesis(14).

The developmental switch from fetal to adult globin synthesis is effectively completed at 14 wk post-term gestation, with HbA representing 95% of the Hb produced after this age(15). However, fetal Hb synthesis is still detectable during the maturation process of individual adult erythroid cells. In studies of erythropoietin-induced erythroid differentiation in vitro, the transition from multipotent progenitors through burstforming units-erythroid and their progeny, colony-forming units-erythroid, and proerythroblast phases is accompanied by a switch from γ- to β-globin expression in individual cells. In these cultures, synthesis of γ-globin occurs predominantly in proerythroblasts and basophilic normoblasts, whereas β-globin synthesis occurs later. Because this pattern mimics the developmental profile in a shorter time frame, it is referred to as the compressed switch(14). It can be postulated that during stress erythropoiesis, immature progenitors are stimulated to enter terminal differentiation more rapidly, causing an accelerated maturation of some erythroid cells. The result is that erythrocytes with increased amounts of HbF appear in the circulation.

An increased level of fetal Hb in erythroid cells is observed in many divergent human conditions. Unlike the pancellular hereditary persistence of HbF syndromes, these conditions do not represent a failure of completion of the erythroid developmental program. but, rather, they are due to an alteration in the coordination of HbF and HbA expression in erythroid precursor cells. The erythropoietic stress associated with increased HbF production as seen in the pathologic conditions included in this study can also occur due to acute blood loss(16), marrow transplantation(17), erythroblastopenia of childhood(16), and severe iron deficiency(18). At a cellular level, studies suggest that this results from the selective recruitment of erythroid progenitors with the capacity to synthesize HbF. At a molecular level, the mechanisms responsible for this increase in fetal Hb synthesis remain unknown.

The data presented in this study show that under conditions of erythropoietic stress, when increased HbF synthesis can occur, there is a very close relationship between the quantification of γ-globin mRNAs as a percent of total non-α-globin mRNAs and γ-globin synthesis as a percent of total non-α-globin synthesis. This is in agreement with a previous study in which a similar correlation was found in normal newborn infants of different gestational ages. The usual time-consuming methods of HbF synthesis determination based on de novo protein synthesis(2) can be replaced by globin mRNAs quantification, by a rapid method that requires only a minimal amount of blood (50 µL).