Main

Because ovine fetal and newborn cardiovascular, pulmonary, and hematologic development resemble human physiology, the sheep is a common model for such studies. Ovine erythropoiesis resembles that of human, and regulation of erythropoiesis and red cell kinetics in the ovine fetus and neonate are currently areas of intense study(1–5). Because the invasiveness of these mechanistic studies preclude similar studies in human infants, these ovine studies are likely to continue. A simple and accurate method for RCV, RCS, or both is critical to optimally investigate neonatal anemia in this animal model.

Recent developments in clinical medicine have implications for uses of such a radioactive method for RCV and RCS. Human recombinant erythropoietin is being evaluated as therapy for anemia of prematurity. Recent advances in obstetrics and fetal medicine permit intrauterine blood transfusion directly into the fetal circulatory system. In both of these areas of investigation and treatment, a method to measure RCV and RCS in the fetus/infant would be useful in understanding the pathophysiology of fetal and neonatal red blood cell disorders and the response to treatment. However, current practical methods expose the fetus/infant to radiation; a method that does not expose the subject to radiation would be particularly useful. To validate such a nonradioactive method in an animal model, a valid radioactive reference method for both RCV and RCS in sheep would be quite useful. For the sheep, we found no such method has been published. Thus, the objectives of this study were to validate the use of [14C]cyanate-labeled red blood cells in sheep to measure both RCV and RCS.

METHODS

These studies were approved by the University of Arkansas for Medical Sciences Animal Care and Use Committee.

Labeling of red cells. The sheep's blood volume was estimated from its weight and the mean blood volume reported for sheep (74.4 mL/kg) by Wade and Sasser(6). Approximately 2% of the sheep's blood was drawn from the external jugular by venipuncture on the same day as the study. Half the volume was drawn into sodium heparin for[14C]cyanate labeling; half was drawn into CPDA-1(7) for 51Cr labeling. Red cells were sedimented by centrifugation at 1000 × g for 15 min at room temperature. The plasma was removed by syringe aspiration using a sterile technique and was saved for later use as described below.

Red cells were labeled with [14C]cyanate, as previously described(8), using the original method of Eschbach et al.(9) except that red cells were washed before labeling. Red cells were washed three times by alternate sedimentation and resuspension in 4 volumes of PBS-glucose (6 mM potassium phosphate, 0.155 M sodium chloride, 22 mM glucose, pH 7.4). [14C]Cyanate (370 mBq/mL (10μCi/mL) of packed red cells, DuPont NEN, Boston, MA) was mixed with the washed red cells and incubated for 30 min at 37 °C in a shaking water bath. The red cells were then washed four times as described above and the final supernatant was removed and discarded, leaving packed red cells.

Red cells were labeled with 51Cr according to the standard protocol adopted by the International Committee for Standardization in Hematology(10). Na51CrO4 (Amersham Corp., Arlington Heights, IL) was mixed with packed red cells at a ratio of 370 mBq/mL (10μCi/mL) packed red cells. The mixture was incubated for 30 min at room temperature. To reduce free 51Cr, ascorbic acid (10 mg of ascorbate/mL of packed cells) was mixed with the red cells and incubated for 30 min at room temperature before infusion into the sheep.

The volumes of the [14C]cyanate and 51Cr-labeled samples were adjusted by addition of the saved plasma to a hematocrit approximately equal to that of the original blood. The [14C]cyanate and 51Cr labels were mixed together in preparation for infusion. Hematocrits of all blood samples were determined by a microcapillary method(11).

Infusion of labeled red cells. Labeled blood was drawn into a sterile disposable syringe. The syringe and blood were weighed. The labeled blood was then infused into the sheep through an i.v. catheter inserted into the external jugular. Infusion of the entire volume of label was accomplished in 1-2 min. The syringe was weighed again to allow determination of the total radioactivity infused. For RCV determination, four to five samples (7 mL each) were drawn between 2 and 30 min after infusion of the labeled red cells. Samples for RCS determination were taken at 1, 2, 4, 7, and 10 d and at weekly intervals thereafter.

14C assay. All samples were assayed in quadruplicate from aliquots weighed to the nearest 0.1 mg; weights were used to calculate concentrations based on the density of ovine blood (1.04 g/mL determined empirically in our laboratory). 14C radioactivity per mL of blood was assayed by the acid acetone method described previously(8). This assay decolorizes globin by extraction of heme using acidification and acetone before scintillation counting of the14 C bound to globin. Volumes used for quantitation were as follows:in vivo circulatory samples at 0.5 mL; normal blood quench control samples at 0.5 mL; and labeled blood samples at 0.05 and 0.1 mL with unlabeled sheep blood added to equalize quench effects. A 0.5-mL blood volume was the maximum that could be decolorized in the assay. The intrassay coefficient of variation for quadruplicate measurements was typically <2%.

51Cr assay. For quantitation of the51 Cr radioactivity, weighing and volume selection were the same as for[14C]cyanate with one exception: a final volume of 1.0 mL was selected rather than 0.5 mL for in vivo and control samples since quench is not inherent to gamma counting. To remove free 51Cr, samples were washed three times using sedimentation at 1500 × g for 10 min at room temperature and resuspension in 2 mL of PBS-glucose. 51Cr radioactivity in the red cell pellet was determined by gamma counting.

Red cell volumes by each label. For each timed sample, the radioactivity per mL of blood was calculated and was plotted versus time. The t = 0 intercept of the linear regression was determined. BV was calculated according to the dilution principle(10) as follows: Equation

AS = total radioactivity infused;CS = radioactivity/mL sample at t = 0;L = radioactivity per g of label; B = background radioactivity per g of unlabeled cells sample; W = weight of label infused (g); S = radioactivity/g mixed venous blood sample;G = specific gravity of blood (g/mL). RCV was calculated as follows: Equation

Post-transfusion recovery at 24 h. PTR24 was calculated as the percentage decline of the concentration of the label in blood at 24 h compared to a reasonable time for mixing after infusion. For these studies, the 5-min postinfusion point was used as follows: Equation

Sequential RCV. Sequential measurement of RCV would be a useful capability to validate. To validate this capability in vivo, we conducted sequential studies in four sheep. In the first sheep, RCV was determined daily for 3 d. Separate aliquots of sheep red cells were labeled with [14C]cyanate and 51Cr and were mixed and stored at 4°C. On the 1st d, a volume of the labeled blood equal to approximately 1% of the estimated BV was infused, and RCV was determined. On the 2nd d, an equal volume of labeled blood was infused, approximately doubling the concentration of circulating radioactivity, and RCV was determined a second time. On the 3rd d, twice the volume of labeled cells was infused (again doubling the concentration of circulating labels), and RCV was determined for a third time. In a similar fashion, agreement between RCV by[14C]cyanate and RCV by 51Cr was determined sequentially in three additional sheep. For two, the interval was 2 d; for the third, the interval was only 20 min.

Multiple determinations of RCV in the same individual over a limited interval of time implies substantial or even accumulating concentrations of residual label. Theoretically, sequential RCV can be determined by correcting for residual radioactivity as follows: Equation where Vi equals the volume of labeled cells infused on the k + 1th infusion, Ca equals the concentration of label on the k + 1th infusion,Ck equals the residual concentration of the label from all previous measurements, and Ck + 1 equals the final concentration of label after mixing of the k + 1th infusion. This equation can be derived for the k + 1th determination of RCV after k previous determinations using successive dilution. This derivation assumes only 1) the conservation of total volume of the labeled red cells over the time for mixing and sampling and 2) no differential sequestration or release of labeled cells. The labeled cells can arise from k + 1th infusion or can be residual from any previous determination.

Statistical analysis. Statistical analyses were performed with StatView software (Abacus Concepts, Berkeley, CA). Agreement between[14C]cyanate and the reference method was assessed by correlation as reflected in the correlation coefficient. Identity was assessed by determining whether the linear regression of paired values in all animals was significantly different from a line of identify as judged by whether the 95% confidence intervals incorporated a slope of 1 and a intercept of 0.

For comparison of RCV by [14C]cyanate to RCV by 51Cr in a individual animal (e.g. sequential volume studies), the RCV values from extrapolation to 0 time were compared using a two-sample t test with pooled variance. Pooled variance was obtained from the variances in the intercept value generated by linear regression(12). The linear regression typically contained four sample times between 2 and 30 min; at each time, at least triplicate measurements of label concentration were made, and the individual measurements were plotted against sample time.

RESULTS

Red cell volume studies. Single RCV determinations. For 19 sheep, RCV determined by the [14C]cyanate method was plotted versus RCV determined by the 51Cr method (Fig. 1). The values agreed almost perfectly (the correlation coefficient = 0.990). The regression line was not significantly different from a line of identity. The slope was 0.94; the 95% confidence interval for slope was 0.87 to 1.00. The y intercept was 40 mL; the 95% confidence interval was -12 to 91. The weights of the 19 sheep studied ranged from 10.9 to 65.9 kg; the RCVs by 51Cr varied from 211 to 1233 mL. The agreement between the two methods was good over the full range of body sizes and commensurate difference in RCV.

Figure 1
figure 1

RCV determined by the [14C]cyanate method vs RCV determined by the 51Cr method for 19 sheep. CC denotes correlation coefficient of the regression line (not shown). Error bars are consistently less than the symbol size.

Full size image

Sequential red cell volume. In determining red cell production, the ability to determine RCV at several intervals that are short compared to the life time of the red cell would be useful. When using a long lived red cell label such as [14C]cyanate, residual and accumulating concentrations of the radioactive label are inevitable. Our equation for correction for the accumulating label indicated that a reasonable signal to noise ratio (i.e. new radioactivity above baseline radioactivity) could be achieved by doubling of the volume of labeled cells infused with each successive measurement. That conclusion was confirmed empirically by several in vitro dilution experiments that explored the characteristics of correcting for baseline radioactivity and background radioactivity (data not shown). RCV was determined in vivo sequentially by both the[14C]cyanate and 51Cr methods (Table 1). In all animals, the agreement between RCV by [14C]cyanate and RCV by51 Cr was well within our estimated analytic variability of coefficient of variance = ±5%.

Table 1 Sequential volume determinations for four sheep
Full size table

Red cell survival studies. PTR24. For the nine animals studied, the mean PTR24 was 83 ± 7% for the[14C]cyanate label and 69 ± 10% for the 51Cr label. This low PTR24 for 51Cr contrasts with the high PTR24 for51 Cr in humans (>96%)(13, 14).

Red cell survival by 51Cr. To test the conclusion of Tucker(15) that 51Cr elutes rapidly from sheep red cells, RCS was assessed in three sheep (Fig. 2). Time to 50% label disappearance(T50) was about 1 wk, providing evidence that 51Cr is unsuitable for RCS studies in sheep.

Figure 2
figure 2

Survival of 51Cr label (n = 3). The51 Cr label leaves the red cell too rapidly to allow determination of RCS.

Full size image

Red cell survival by [14C]cyanate. A single survival curve of the [14C]cyanate label is depicted in Figure 3 (sheep 2) to emphasize its curvilinear nature. For this sheep (and for each sheep studied), a two-component model adequately fits the survival data. The portion of cells in the two components and time parameters of the components vary between sheep. In the first three sheep studied (Fig. 4A), early loss was substantial. The exponential component accounted for as little as 11% but as much as 48% of the total cells. Half-lives of these exponential components varied from as short as 4 d to as long as 16 d. Despite these early losses, the MPLs obtained from the extrapolation of the linear component averaged 118 d.

Figure 3
figure 3

Survival of [14C]cyanate labeled red cells in sheep 2. A two-component model fits the survival curve well and allows determination of the MPL.

Full size image
Figure 4
figure 4

Survival of [14C]cyanate labeled red cells.(A) Early experiments showing nonlinear label decline (n = 3). (B) Late experiments showing linear label decline (n = 6).

Full size image

Empirical modifications were made to maximize PTR24 and minimize the early exponential loss. For example, incubation time for labeling was reduced from 1 h to 30 min; red cells were equilibrated with plasma for 30 min before reinfusion. Using these modifications as specified in “Methods,” the linear component was increased to greater than 95% of the total (Fig. 4B), and an exponential component was no longer necessary for a correlation coefficient of >0.99. The average MPL observed in the six sheep was 114 ± 12 d (Fig. 4B). These values agreed well with the range of 70-153 d reported by Tucker using59 Fe pulse-chase(15).

DISCUSSION

To our knowledge, the studies presented here are the first application of[14C]cyanate labeling for RCV determination in any species and the first application of the technique for RCS determination in sheep. The validation of this method adds another tool for investigation of ovine erythropoiesis and red blood cell kinetics, which is an important model for studying anemia in human infants.

RCV. The dilution method using red cells labeled with51 Cr is the reference method (the “gold standard”) for RCV in sheep(11). Accordingly, 51Cr was used as the reference method in the RCV studies presented here. Assuming that51 Cr-labeled red cells provided a suitable reference value, this study provides evidence that [14C]cyanate labeling of red cells can be used to determine RCV accurately and precisely. Within experimental error, the results of the two methods are identical. For a simple determination of RCV, the [14C]cyanate method offers no theoretical advantages; the costs in materials and effort were similar.

Normalized for body weight, the RCV results obtained here also agreed reasonably well with RCV in sheep published by Wade and Sasser(6), even though those investigators studied sheep over a more narrow weight range. Using autologous red cells labeled with both59 Fe and 51Cr, Wade and Sasser reported RCV of 21.1 ± 2.2 mL/kg (mean ± 1 SD) for a group of 40 sheep having a mean weight of 32.6 ± 3.6 kg. For the sheep studied here, the mean RCV was 18.8± 3.9 mL/kg (mean ± 1 SD) determined by the 51Cr method and 18.8 ± 3.9 mL/kg (mean ± 1 SD) determined by the[14C]cyanate method. This group of 19 sheep had a similar mean weight of 36.6 ± 14.7 kg, but much wider range of weights compared to Wade and Sasser's group of 40 sheep.

In this study, the mean BV was 63.5 ± 8.7 mL/kg for 51Cr and 63.6 ± 9.6 mL/kg for [14C]cyanate, whereas Wade and Sasser reported a mean BV of 74.4 ± 8.7 mL/kg. Thus, the agreement of blood volumes normalized by body weight with the literature was not as good as for RCV. These differences may relate to the different populations of sheep studied or to study conditions; our observations suggest that plasma volume equilibration and sequestration of red cells occur in minutes to hours in the sheep and relate to the handling and state of excitement of the sheep at the time of study. The design of these RCV studies deliberately did not address the physiology of circulation in the sheep (e.g. splenic sequestrations). Limitations and caveats appropriate to determination of RCV by 51Cr likely apply as well to RCV determination by[14C]cyanate. An example of such limitations is provided by the data of sheep 3 in the sequential volume studies. As assessed by either label, RCV increased by about 12% over 2 d (Table 1). This increase in RCV could not have arisen from new synthesis, which replaces about 1% of the circulating red cell mass per day at steady state in sheep that are not anemic. The 12% increase in detectable RCV suggests that some splenic sequestration of unlabeled cells might have occurred on the 1st d before infusion of labeled red cells; if release occurred later, a greater circulating RCV would be observed by the second measurement. In accord with the possibility of sequestration is the observation that RCV was unusually small (11.8 mL/kg) for the first measurement.

PTR24. These in vivo studies provide evidence that the blood concentration of 51Cr at 24 h averaged 69% of the initial value. This reduction in concentration could reflect rapid elution of 51Cr from sheep cells, or removal of the sheep cells from circulation, or both. Data from this study do not directly address these two mechanisms, but we suspect that the decrease reflects the first phase of a rapid diffusion of 51Cr from sheep red cells and that the diffusion continues in the following days. A similar rapid loss of 51Cr from canine cells over the first 24 h and over the next few weeks has been reported(16). In contrast, PTR24 of >96% have been observed in human studies(13, 14), and thereafter the elution rate of 51Cr from human RBC is only about 1% per day(13).

The greater PTR24 for [14C]cyanate labeled cells compared with 51Cr (83% versus 69%) could be the result of any one or a combination of the following three processes: 1) less rapid14 C label loss from red cells, 2) smaller losses of red cells that are injured and removed from circulation, and 3) labeling of the intracellular proteins (hemoglobin) by unreacted [14C]cyanate, which likely continues in vivo for at least 24 h. The design of our studies does not permit assessing the relative contributions of these phenomena, but there is likely an important contribution due to less rapid label loss. Original studies in other species(16, 17) and our studies in the sheep suggest that [14C]cyanate is a permanent red cell label by virtue of its covalent carbamylation of the terminal amino group of hemoglobin.

RCS. Assuming that [14C]cyanate is indeed a permanent red cell label, the concentration of radioactivity per volume of blood (or per volume of red cells) reflects true survival of the red cells. After we made the labeling process less harsh, RCS curves were linear. This linear decline is consistent with uniform labeling of a mixed population of red cells containing equal numbers of cells of different ages, which were labeled at equal densities. We propose that the exponential component seen in sheep 1-3 represents early loss of cells damaged by labeling and handling, whereas the linear component represents the survival of relatively undamaged cells and extrapolates to an approximately normal MPL.

We followed survival in one sheep for 3 mo; we found no evidence of reincorporation of 14C into red cells (Fig. 3). Given these decay characteristics and the agreement between these values and those of Tucker(15) using 59Fe pulse/chase labeling, it seems likely that the linear portion of the RCS of[14C]cyanate approximates the physiologic life span of the red cell. However, a modest reduction of MPL or time to 50% survival (e.g. 10%) would likely not be detected by these studies simply by virtue of the analytical precision (coefficient of variance = ±5%).

Estimates of MPL were not corrected for somatic growth of the sheep studied. Average percent body growth over the time interval of the RCS studies was 55%; the range was 2 to 150%. We elected not to correct for somatic growth because it would preclude comparison of our results to those of Tucker(15) who did not correct for growth. However, it must be remembered that accurate absolute estimates of red cell life span would require correction for the expansion of the vascular volume.

The [14C]cyanate method offers some advantages over the 59Fe pulse/chase method and the differential hemolysis method-those previously used to measure red cell life span in the sheep(15). The[14C]cyanate method avoids the inaccuracy due to 59Fe reincorporation into red cells observed with the 59Fe pulse/chase method. 59Fe pulse/chase method also requires assessing survival of the label for about 90% of the MPL of the red cell; in contrast an early linear portion of the [14C]cyanate label RCS curve can be used. Moreover,59 Fe pulse/chase does not allow RCV determination. Other methods to permanently label red cells have been studied (e.g. diisopropyl[32P]phosphofluoridate) and glycine ([15N]glycine and[14C]glycine). They have potential advantages over 59Fe pulse/chase but are somewhat impractical and have not been validated in sheep(18). Differential hemolysis has been used in the sheep(15). The [14C]cyanate method is technically easier to perform than the differential hemolysis methods and does not place restrictions on use of heterologous red cells.

Summary. The data of this study provide evidence that circulating RCV can be accurately measured in sheep using[14C]cyanate-labeled red cells. Although use of very short lived isotopes may offer a superior approach to sequential determination of RCV, this study provides evidence that RCV can be determined sequentially in the same sheep with this long lived isotope. This study also provides evidence that RCS can be accurately determined in sheep using [14C]cyanate. Survival of the [14C]cyanate label over the first 24 h in circulation(PTR24) is likely the result of complex processes involving both cell loss due to damage and continued in vivo labeling. Accordingly, RCS should be referenced to d 1 or later. With proper handling of red cells, damage due to labeling can be minimized and the expected linear decline is observed, suggesting that survival of [14C]cyanate-labeled red cells approaches the true physiologic life span. Eschbach et al.(9) have applied [14C]cyanate labeling to measure RCS in normal and uremic subjects. They concluded that [14C]cyanate is a permanent red cell label.

Unfortunately, the long half-life of [14C]cyanate might preclude the use of this method in children. Although the radiation delivered by the shorter lived 51Cr has never been shown to produce harm at recommended doses, a completely nonradioactive method would be desirable. Methods have been published that use stable isotopes with subsequent neutron activation(19) or x-ray fluorescence(20), but these methods require highly specialized equipment(19, 20), are not sufficiently sensitive for the small volumes of blood available from neonates(20), or have not been evaluated as RCS markers(19, 20). Accordingly, none of these methods appears to be widely used. Recently, we have used this [14C]cyanate method as the reference method to validate a nonradioactive method for RCV and RCS in sheep. In turn, this nonradioactive method is being used to measure RCV and RCS in neonates.