Maternal, fetal, and amniotic fluid transport of thyroxine, triiodothyronine, and iodide in sheep: a kinetic model.

0O13-7227/78/1032-0567$02.0O/O Endocrinology Copyright © 1978 by The Endocrine Society

Vol. 103, No. 2

Printed in U.S.A.

Maternal, Fetal, and Amniotic Fluid Transport of Thyroxine, Triiodothyronine, and Iodide in Sheep: A Kinetic Model* ROBERT A. McGUIRE AND MONES BERMAN Laboratory of Theoretical Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20014 ABSTRACT. A mathematical model of iodine kinetics in maternal and fetal sheep has been developed by combining separate iodide, T3, and T4 subsystems. The individual subsystem models were developed from literature studies of maternal-fetal exchange under thyroid-blocked and unblocked conditions. Rates of exchange, concentrations, and spaces of distribution were calculated by the SAAM computer program. The models for each of the subsystems required exchange compartments within the mother and fetus, exchanges between maternal and fetal circulations, and between the fetus and amniotic fluid. The fetal-amniotic fluid exchange was observed directly for iodide and indirectly for T3 and T4. No exchange between mother and am-

S

TUDIES of thyroid hormone metabolism have indicated that in man and sheep the fetal thyroid system functions independently of the mother during late gestation (1) and may contribute to fetal hormone economy in early fetal life as well (2). Although both T4 and T 3 are present in the fetus during late gestation, fetal T3 levels are significantly lower than those found in neonates and adults (3), whereas T4 levels are similar. The decreased fetal T 3 concentration must involve either reduced production or increased catabolism in the fetus, preferential fetus-to-mother transport, or a combination of these. Dussault et al. (4) have demonstrated an exchange of T3 in sheep between mother and fetus which is a major component of fetal T3 input and loss, with a net transfer to the fetus greater than 0.8 ng/day, accounting for a least xk of the total fetal loss. Received October 13, 1977. Address requests for reprints to: Dr. Robert A. McGuire, General Medical Research Service, Veterans Administration Hospital, Washington, DC 20422. * Portions of this work were presented at the Annual Meeting of the Endocrine Society in June, 1975, New York, New York.

niotic fluid was required. It is possible that the amniotic fluid acts as a reservoir for these and other substances. Maternal-fetal kinetics suggest that low fetal T3 levels are maintained by an active transport of T 3 from fetus to mother, a decreased transport from mother to fetus, and a low fetal T3 production. The model also requires that all fetal T3 loss occur via transport to the maternal system rather than via fetal utilization. In contrast, the fetal T4 system is largely autonomous, the small maternal exchange not significantly contributing to the fetal T4 economy. Fetal iodide seems to be supplied by a facilitated bidirectional exchange with the mother. (Endocrinology 103: 567, 1978)

The role of T4 exchange is less clear. Robin et al. (5) reported a small but significant transport of T4 between fetus and mother, and Dussault et al. (6) reported that the T4 exchange contributes less than 2-3% of total fetal supply and is not important in maintaining fetal T4 levels. Fetal production of T4 requires an iodide (I~) supply from the mother, and a considerable iodide exchange between mother and fetus has been demonstrated by Robin et al. (5). The role that amniotic fluid plays in fetal thyroid hormone economy has not been clearly defined. Hollingsworth and Austin (7) reported the presence of T4 in human amniotic fluid at term, but the mechanisms by which it appears and is removed are not known. In sheep, tracer iodide has been injected into amniotic fluid as well as fetal and maternal blood by Book et al. (8) and the level of tracer in each site was followed. The results indicated a significant exchange between all three sites, but the routes of exchange were not resolved. Although considerable kinetic data are available on the exchange of T4, T3, and I",

567

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 16 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.

McGUIRE AND BERMAN

568

few efforts have been made to integrate these into a unified model. Dussault and coworkers (4) utilized single compartment maternal and fetal models to estimate the transfer of T3, but did not take into account the distribution kinetics within the maternal and fetal subsystems or the possible role of the amniotic fluid. In order to integrate the available data and to elucidate the mechanisms responsible for the various responses, a comprehensive model that accounts for I~, T3, and T4 kinetics is presented here.

200 DC

Endo • 1978 Vol 103 • No 2

DRTfl FROM" DUSSRULT ET RL. ENDOCRINOLOGY 90- 1301,, 1972 + MRTERNRL INJECTION X FETRL :INJECTION

CO O

a

Model Development and Results Modeling and least squares fitting of the data were performed on Univac 1108 and Digital Equipment Corp. PDP-10 computers using the SAAM program (9, 10). Compartmental models were developed and used to reproduce each reported experiment. Reasonable assumptions were introduced where necessary until all experimentally observed features could be accounted for (11, 12). The experimental data used in the development of the model came from the literature and were collected under two different conditions. Some studies were carried out with iodide or perchlorate as thyroid-blocking agents while others were carried out without blocking agents. The composite model is an integration of separately derived models for T4, T3, and I" subsystems in the mother, fetus, and amniotic fluid. T3 subsystem (blocked) The T 3 model was developed primarily from the tracer experiments of Dussault et al. (4) and from the data of Erenberg et al. (1), in which the thyroid was blocked with perchlorate. In these experiments, tracer [125I]T3 and [131I]T3, respectively, were administered simultaneously to mothers and their fetuses at 0.8-0.9 gestation (gestation is 150 days in the sheep), and fetal and maternal plasma radioactivities were measured for the following 24 h (Fig. 1). The derived model is shown in Fig. 2 and the steady state values are given in Table 1. The maternal T 3 subsystem is represented by two compartments (C3 and C4).

TIME - HOURS FIG. 1. T3 tracer data. In this figure and those which follow, symbols represent observed data and solid lines represent data generated in the model.

FIG. 2. T3 model under blocked conditions. In this figure and those which follow, circles represent compartments and are numbered for identification. Arrows represent rate constants (LI,J) defined as the fraction of compartment J flowing to compartment I per U time. Arrows not leading to another compartment represent irreversible loss pathways from the system. Rectangles represent delays and the double arrow represents the steady state input. Triangles represent observed quantities which are linear combinations of one or more compartments.

Compartment 3, which received the initial tracer injection, has a PEV1 of 6.4 liters or approximately 13% BW (Fig. 2). This volume probably includes rapidly equilibrating tissues as well as plasma. C4 in exchange with C3 has a PEV of 19.9 liters (41% BW) and probably includes both extra- and intracellular volumes. 1 The following abbreviations were used: PEV, plasma equivalent volume; Cm, compartment m in model; (LI,J), the fraction of compartment J transferred to compartment I per unit time. When 1 = 0, the fraction of J lost from the system.


T4, T3, AND T KINETICS DURING GESTATION TABLE 1. T 3 subsystem under blocked conditions: steady state values

BW (kg) Total T 3 (/xg) Total PEV (liters) Concentration (jug/liter) Model Observed a 6

Maternal

Fetal

Amniotic fluid

48.0 20.8 26.3

1.97 0.43 2.39

0.72

0.18" 0.18

0.36* 0.20

0.79° 0.79

Fixed to observed value. Based on estimated volume of 2 liters.

Together the volumes approximate total body water and are comparable to those reported by Belshaw et al. (13) for the dog. The fetal T 3 sybsystem is represented by two compartments, C23 and C24. Although C24 was required by the fetal data, it was not possible to resolve uniquely the size of this compartment; therefore, it was set so that the ratio of C24:C23 in the fetus would equal the C4:C3 ratio in the mother. Additional data will be required to test the validity of this assumption. C25 was introduced to account for an observed slow phase in the fetal T 3 kinetics, due perhaps of amniotic fluid. A 20min delay was also needed to account for an observed delay in the appearance of radioactivity in the mother when tracer T3 was administered to the fetus. No delay from mother to fetus could be resolved from the available data. The PEV of C23 was approximately 0.581 or 29% BW, and that of C23 combined with C24 was 2.39 liters, or 119% BW. The larger relative size of the plasma compartment in the fetus was also observed for the T4 and I~ subsystems and is probably due to the inclusion of placental spaces. The slow compartment (C25) required in the fetal system was not seen in the maternal system. The large size (4 liters PEV) and slow decay (t^ = 13 h) of this compartment suggested that it might be amniotic fluid. The routes of fetal to amniotic fluid T 3 losses have not been identified. Reabsorption from swallowed fluid probably provides for the return of T 3 to the fetus (14). A direct exchange between mother and amniotic fluid was not

569

required but cannot be excluded with the present data. Assuming C25 is amniotic fluid, its T 3 concentration in sheep is estimated from the model to be about 36 ng/dl, based on an estimated amniotic fluid volume of 2 liters. Measurements by Burman and Wartofsky have shown amniotic fluid T3 levels of 25 ng/dl in human term pregnancies (15) and 20 ng/dl in sheep (Burman, K. D., and L. Wartofsky, unpublished observations) sampled between 110-135 days gestation (0.8 gestation). T4 subsystem (blocked) The T4 model was derived primarily from the tracer data of Robin et al. (5), and from the data of Erenberg et al. (1). In these experiments [131I]T4 and [125I]T4 were given at 0.8 gestation to mothers and fetuses, respectively, and radioactivities in both fetal and maternal blood were measured for periods up to 192 h postinjection. The thyroids were blocked with potassium iodide. The observed and modelpredicted responses are shown in Fig. 3, and the model is shown in Fig. 4. Steady state values for this system are given in Table 2. The data shown are the mean of six animals, except for the last four points when only two 1000

DRTR FROM- ROBIN ET RL. HORMONES 3- 235. 1972

CC LLJ

+ MRTERNRL INJECTION X FETfll INJECTION CO

o Q

a LLJ

I—

.01 0

50

112121

150

2121(9

TIME - HOURS FIG. 3. T4 tracer data. , Best fit obtained without the postulated contaminant. The data are the mean of six animals, except for the last four points which represent one or two animals.


McGUIRE AND BERMAN

570

FIG. 4. T4 model under blocked conditions. See description under Fig. 2. Fetal data are the sum of fetal plasma T< (Cll) and contaminant (CIO). TABLE 2. T4 subsystem under blocked conditions: steady state values

BW (kg) Total T4 (/xg) Total PEV (liters) Concentration (jug/liter) Model Observed

Maternal

Fetal

Amniotic fluid

55.7 292.0 5.83

1.35 29.5 0.49

18.4

50.0" 50.0

60.0" 60.0

9.2* 5.0

" Fixed to observed value. 6 Based on estimated volume of 2 liters.

animals survived. The maternal system was represented by a two-compartment model. The plasma-containing compartment (Cl) has a PEV of 2.7 liters (4.8% BW) and is considerably smaller than the plasma-containing compartment of the T3 system. The total plasma equivalent volume is 11% BW and compares well with the 13 and 15% observed in dog and man (13, 11). The values of the exchange rates between Cl and C2 (L2,i,Li,2; Fig. 4) are comparable to those found in man but slow compared to the dog. The fetal T4 subsystem required a two-compartment model similar to the T3 subsystem. The PEV of the plasma-containing compartment (Cll) was 17% BW and the total fetal PEV was 37%. As in the T 3 subsystem, a portion of this is probably due to the plasma in the placenta. A delay from fetus to mother was introduced to fit the early maternal response to the fetal dose. The delay is about 1 h but cannot be accurately determined.

Endo Vol 103

1978 No 2

The slow fetal exchange compartment seen in the T3 sybsystem was also required for T4. If this compartment is due to amniotic fluid, the model predicts an amniotic fluid concentration over 10 times that observed by Burman and Wartofsky [(Ref. 15 and unpublished observations)] and by Chopra (16). Since this discrepancy could not be eliminated without seriously compromising the fit and because a similar slow phase was not seen in the data for the mother, we attributed part of the slow phase to a small contaminant with an assumed half-life of 7 days. The long half-life required makes it unlikely that the material represents rT3 or other T4 degradation products. It is also possible that this could represent a measurement artifact, due to the very low levels of T4 transfered. Under this assumption, about 1.5% of the initial dose as a contaminant was adequate to explain a steady state T4 concentration of 0.5 jug/dl in the amniotic fluid, as observed by Burman and Wartofsky. The dashed line in Fig. 3 represents the best fit of the fetal plasma data obtained without the contaminant. If a contaminant were present it would also appear in the maternal system; however, its appearance in the mother would be masked by the higher T4 tracer levels appearing in the mother after fetal injection and by the slower turnover of the maternal T4 system. Additional T4 kinetic data woul be helpful to resolve this problem. T4 to T3 conversion An estimate of T4 to T3 conversion in the mother was derived from the data of Fisher et al. (17). In that study, unlabeled T4 was administered to nonpregnant adult sheep and the subsequent levels of T4 and T 3 in plasma were followed. This design provides a lower limit on conversion, as conversion sites which do not exchange with plasma are not seen. Using our maternal T4 and T3 models, we simulated the experiments by diverting part of the T4 loss pathway from the plasma compartment to the T3 plasma compartment. The fit shown in Fig. 5 was obtained by diverting 3.6% of the total T4 loss and accounts for 77% of the total T3 production. In contrast, Fisher


T4, T3) AND I" KINETICS DURING GESTATION 1006)

DflTfl FROMi FISHER ET flL. ENDOCRINOLOGY 91 • 11(11. 1972

50

HOURS

FIG. 5. T4 to T3 conversion in adult sheep under blocked conditions.

et al. (17), using a single compartment analysis, calculated a conversion of almost 10% of the T4, supplying 91% of T3. Alternate conversion pathways via the exchange compartments were also tested, but each was discarded as it introduced too large a delay in the appearance of plasma T3. Thus, it seems that the sites of T4 to T3 conversion are in rapid equilibrium with plasma. This is consistent with the recent findings of Chopra (18) that T4 to T3 conversion occurs primarily in rat liver and kidney homogenates in an in vitro system. Since data comparable to those used in the adult were not available for the fetus, it is only possible to estimate a maximum T4 to T3 conversion. If all fetal T3 were derived from T4, about 3% of fetal T4 would be converted. Any release of T 3 from the fetal thyroid or intake from the mother would further reduce T3 conversion.

571

The total space, 50% BW, is comparable to that found in the dog (13) and somewhat higher than in man [35% BW (11)]. A two-compartment model was sufficient to describe the fetal I" data. Each compartment is approximately 85% BW and together comprise three times the total maternal PEV on a percent body weight basis. The location of these I" stores in the fetus is not known, but it is likely that placental tissue contributes significantly to the total space. Although no amniotic fluid compartments were necessary to fit the blocked iodide kinetics, such a subsystem was introduced in light of the subsequent development of the nonblocked iodide model (see Fig. 9) for which amniotic fluid data were available. The steady state amniotic fluid iodide concentration required a much smaller iodide compartment than that required in the unblocked model. 108

J

DflTfl FROM- ROBIN ET flL. HORMONES 3' 235. 1972

10 IT) O

a a UJ

10-j

X FETRL INJECTION + MRTERNRL INJECTION FETRL PLRSMR MflTERNRL PLRSMR

+

a o 0

24

48

72

96

120

144

TIME - HOURS FIG. 6. Iodide tracer data under blocked conditions.

Iodide subsystem (blocked)

X

The data of Robin et al. (5) shown in Fig. 6 J V were used to derive the iodide for mother and fetus shown in Fig. 7. Steady state values for the subsystem and for the unblocked model are given in Table 3. The experiments were carried out under conditions similar to those used in the T4 studies. A two-compartment model was required for the maternal subsysAMNIOTIC FLUID tem. The plasma-containing compartment has a PEV of 11.5 liters (21% BW), and is in rapid FIG. 7. Iodide model under blocked conditions. See deexchange with an additional PEV of 15.8 liters. scription under Fig. 2. .O32/hr 1.117 mg/hr)

C15


572

McGUIRE AND BERMAN

Endo i 1978 Vol 103 i No 2

TABLE 3. Iodide subsystems: steady state data Unblocked

BW (kg) Total iodide (fig) Total PEV (liters) Iodide concentration 0 (jug/liter) Model Observed T 4 concentration (/xg/liter) Model Observed

Blocked"

Maternal

Fetal

AF*

Maternal

Fetal

AF

67.0 191.0 60.0

3.3 86.3 12.0

39.6 1.3

55.7 8653.0 27.3

1.35 1086.0 2.33

59.2

371.0 350.0

440.0 540.0

31.0rf 31.0

506.0 540.0

35.6P 31.0

49.0 50.0

60.0 60.0

4.7 5.0

350.0rf 350.0

" Does not include blocking iodide. 6 Amniotic fluid. c Includes T 4 (and slow iodine under unblocked). d Fixed to observed value. e Based on assumed AF volume of 2 liters. DRTR FROM- ROBIN ET RL.

Total blocked system The T4, T3, and I~ subsystems were interconnected and used to fit the total blood data of Robin et al. (5) obtained after labeled T4 injections (Fig. 8). In the combined model, approximately half of the T4 and T 3 loss pathways was diverted to iodide; the remaining losses represented nonrecycled fecal losses. A T4 to T3 conversion pathway was included in the mother. No T4 and T3 synthesis paths were incorporated, as the data were obtained under blocked conditions. Total activity was represented as the sum of plasma T4, T3, and I" activities. Iodide subsystem (unblocked) A model for iodide distribution under unblocked conditions was constructed using the data of Book et al. (8, 19). 131F was administered to the mother and 125I~ was injected into either the fetus or the amniotic fluid in a total of five animals. The plasma of the mother and fetus and the amniotic fluid were sampled over a 48-h period. These studies provided the only direct evidence for amniotic fluid exchange. The observed and calculated data are shown in Fig. 9 and the model is shown in Fig. 10. The structure for the fetal and maternal I" subsystems is similar to that described for blocked conditions. Amniotic fluid I~ is represented by a two-compartment subsystem. The second compartment may be due to temporary blockage of part of the fluid by the

HORMONES 3- 235. 1972

50

100

150

20(21

TIME - HOURS FIG. 8. Total blood radioactivity after T4 injection. The activity is represented as the sum of T4, T3, and iodide from the individual models.

fetus, as it appeared in only one of the two animals studied. Small delays (1 and 2 h) were required in the fetal-amniotic fluid exchange pathways. As in the blocked iodide model, no delay was necessary between the fetal and maternal subsystems. It was not possible to fit the late phase of the unblocked data using only the blocked I" model. Since under unblocked conditions significant hormonal iodine could be produced during the 48-h study, simplified single compartment thyroid and T4 subsystems were added to both the maternal and fetal models. A T3 subsystem was not included because it would not contribute significantly to the total


T4, T3) AND r KINETICS DURING GESTATION

573

MRTERNRL INJECTION HMNIOTIC FLUID

*

DflTfl FROM' SOOK ET fiL. HERLTH PHYSICS 26- 533. 1974

.01

01

m

FIG. 10. Iodide model under unblocked conditions. See description under Fig. 2. C1-C6 represent iodide; C14 and C7 represent thyroid; and C15, C8, and C9 represent T4. The slow iodinated material is in C17 and C18. The triangles represent the calculated data, which are the weighted sum of the iodide, T4, and slow I compartments shown.

FETRL INJECTION

en a a a UJ

Q

O

RMNIOTIC FLUID INJECTION

12

24

36

4b

TIME - HOURS FIG. 9. Iodide tracer data under unblocked conditions.

radioactivity measured. The amniotic fluid T4 compartment was comparable to that in the blocked T4 model except that its volume was constrained to the total amniotic fluid volume, as determined from the I" data.

An additional problem appeared in the steady state solution of the unblocked I~ system. The ratio of amniotic fluid to maternal or fetal total iodine concentration in the model was much higher than observed. Since the parameters for fetal-amniotic fluid exchange of I~ are well defined by the tracer data and the concentrations of T4 in each site were known, this discrepancy could not be due to either iodide or T4. Therefore, we elected to introduce slowly exchanging compartments C17 and C18, which matched the observed steady state concentrations but do not contribute significantly to the tracer data. The physiological identity of these pools is uncertain, but it is possible that they represent iodinated proteins. Also, since tracer and steady state data were not collected in the same animals, it is likely that at least some of the discrepancy represents animal and experimental variations. Maternal-fetal and fetal-amniotic fluid clearance values for each of the substances studied are tabulated in Table 4. The calculated clearance rates of amniotic fluid for T3, T4, and I" under blocked conditions are based on an assumed fluid volume of 2 liters. For T4 and T3, the clearance rates between mother, fetus, and amniotic fluid were assumed to be the same as those obtained under blocked conditions.


574

McGUIRE AND BERMAN

TABLE 4. Clearance rates (liter/h) M-»F

F-»M

F—AF AF"^F 0.17 0.04 0.004 0.05 0.31 0.15 0.0023 0.0140

1.2 Iodide (unblocked) 2.7 Iodide (blocked) 0.37 0.25 T3 (blocked) 0.06 0.41 T< (blocked) 0.0005 0.0040 0.2 0.2 0.2 Water M, mother; F, fetus; AF, amniotic fluid. " Based on assumed AF volume of 2 liters.

0.2

Discussion The bulk of fetal T3 seems to be derived via transport from the mother, based on the amount of maternally administered labeled T3 appearing in the fetal circulation. This pathway alone accounts for a steady state fetal T3 concentration of 14.8 ng/dl. Since the observed fetal T3 in the Dussault study was estimated to be between 8 and 18 ng/dl, little could be derived from endogenous sources, either via fetal T3 synthesis or T4 to T 3 conversion. The low T3 production from T4 might be compensated by a bigger production of rT3, as suggested by Chopra et al. (20), such that the total T4 economy remains unchanged. It is also possible that some T4 is converted to T3 at tissue sites which do not release T3 to the circulation before eventual breakdown. This "invisible" T 3 production would not contribute to the plasma T3 levels. In contrast to our results, Chopra et al. (20) have calculated that the bulk of fetal T 3 is endogenously derived. This is based on the levels of T 3 in the fetal thyroid and on the assumption that the hormones are secreted from the gland in proportion to their content. Further work will be needed to resolve this point. The T3 that does appear in fetal plasma is removed by transport to the mother, rather than by local utilization. We are forced to this conclusion by the high maternal tracer levels seen after fetal T3 injection. There is fair agreement between our results and those of Dussault et al. (4) when the amniotic fluid is included as part of the fetal system. The major discrepancies arise in the steady state T 3 utilization in both mother and fetus and the transport of T 3 from mother to fetus. Dussault et al. (4) calculate utilization rates which are 50% higher and transport rates

Endo i 1978 Vol 103 , No 2

of T3 from fetus to mother which are 30% lower than ours. They calculate a net transport of T3 to the fetus while our model predicts a net transport from the fetus for fetal concentrations of 14.8 ng/dl and higher. Erenberg et al. (1) obtained a fractional rate constant from fetus to mother of 0.007/h based on a single compartment. Conversion of our constant (0.7/h), which represents the fraction of the plasma compartment (C23) transported per hour, to that form yields a value of 0.050/h, based on the data of Dussault et al. (4), and 0.013/h, based on the data of Erenberg et al. (1). This difference of a factor of 4 is due mostly to the differences in the levels of labeled maternal T 3 after injection of labeled T3 in the fetus seen in these experiments. Water exchange between maternal and fetal systems was assumed to be about 0.21/h in both directions, based on the human data of Hutchinson et al. (21). The daily I" clearance under blocked conditions (Table 4) is close to this estimate, and it is likely that it is accomplished by similar bulk transport and diffusion processes. At lower iodide levels, during unblocked conditions, it seems that clearance is facilitated in both directions, with an 8-fold increase in maternal to fetal clearance compared to a 5-fold increase in fetal to maternal clearance. In vitro studies of placental iodide transport also indicate that active processes are involved (22). The contribution of this mechanism at higher iodide concentrations may account for the somewhat elevated values observed when compared to the estimates for water. As previously observed by other workers (4, 5), T4 clearances are low. It also seems that T 3 clearance from mother to fetus is similarly low. It is possible that in each case diffusion of free hormone is unimpaired but that exposure time of the circulating fetal or maternal blood to placental exchange sites is insufficient for significant amounts of bound hormone to be released. The higher percentage of unbound T3 compared with T4 might then account for the larger T3 maternal to fetal clearance value. This mechanism, however, would not be sufficient to account for the high fetal to maternal clearance seen, which is 60% higher than that seen for unblocked iodide. It


T4, T3, AND I" KINETICS DURING GESTATION seems that an active mechanism is specifically involved in the transport of T 3 from the fetus, which would thus account for the low fetal plasma levels seen. The model supports the observations of Hollingsworth and others that T4 and T 3 are present in amniotic fluid (7, 12, 14). Whether these hormones derive from fetus, mother, or both is not known, although Sack et al. (23) have suggested that exchange with both must be involved due to the lack of correlation of amniotic fluid levels with either site alone. We have not included a direct maternal-amniotic fluid pathway, however, as it was not required by any of the available data, and could not be supported by the iodide data of Book et al. (8, 19). The fetal-amniotic fluid pathway was required in the iodide model and was suggested by the late phase of the fetal T4 and T3 data. The probability that T3 and T4 are delivered to the amniotic fluid by biliary excretion into meconium is supported by the observed fecal excretions of endogenous T4 in the dog and human. However, significant meconium excretion in utero has not been demonstrated, and its presence at birth is regarded as a sign of possible fetal distress (24). If T4 and T3 are excreted via this route, then there should be a delay between removal of hormone from the fetal circulation and its appearance in amniotic fluid. A delay of about 1 h was required for the iodide of Book, which is probably more consistent with a urinary excretion path for that ion. Reabsorption via swallowing would also be subject to a delay, as was seen in the iodide data. In view of the slow turnover of amniotic fluid, it is likely that significant concentrations of even poorly excreted materials would build up in amniotic fluid with time. Using an estimated amniotic fluid volume of about 2 liters and a swallowing rate of 200 ml/h (21), about 10% of the amniotic fluid pool would be swallowed per hour. The model estimate for T 3 absorption from amniotic fluid is 0.077/h. Thus, it seems that about 75% of swallowed T3 is absorbed on each pass through the intestinal tract. Similar calculations for I~ and T4 yield estimates of about 50 and 7%, respectively. Whether this discrepancy is due to preferential T4 binding to amniotic fluid proteins or reflects a less efficient absorption

575

of T4 in the fetal intestine is not known. Although it is increasingly apparent that rT 3 plays an important role as a less active metabolite of T4 in both the mother and the fetus, we have not incorporated a rT3 model due to the lack of suitable kinetic data available in the literature. We plan to expand the model as such data become available. References 1. Erenberg, A., K. Omori, W. Oh, and D. A. Fisher, The effect of fetal thyroidectomy on thyroid hormone metabolism in maternal and fetal sheep, Pediatr Res 7: 870, 1973. 2. Fisher, D. A., J. H. Dussault, A. Erenberg, and R. W. Lam, Thyroxine and triiodothyronine metabolism in maternal and fetal sheep, Pediatr Res 6: 894, 1972. 3. Larsen, P. R., Direct immunoassay of triiodothyronine in human serum, J Clin Invest 51: 1939, 1972. 4. Dussault, J. H., C. J. Hobel, J. J. DiStefano, III, A. Erenberg, and D. A. Fisher, Triiodothyronine turnover in maternal and fetal sheep, Endocrinology 90: 1301, 1972. 5. Robin, N. I., H. A. Selenkow, V. S. Fang, S. Refetoff, G. Piasecki, H. Rauschecker, and B. T. Jackson, Bidirectional thyroxine exchange in pregnant sheep, Hormones 3: 235, 1972. 6. Dussault, J. H., C. J. Hobel, and D. A. Fisher, Maternal and fetal thyroxine secretion during pregnancy in the sheep, Endocrinology 88: 47, 1971. 7. Hollingsworth, D. R., and E. Austin, Thyroxine derivatives in amniotic fluid, J Pediatr 79: 923, 1971. 8. Book, S. A., H. G. Wolf, H. R. Parker, and L. K. Bustad, The exchange of radioiodine in pregnant and fetal sheep, Health Phys 26: 533, 1974. 9. Berman, M., E. Shahn, and M. F. Weiss, The routine fitting of kinetic data to models: a mathematical formalism for digital computers, Biophys J 2: 275, 1962. 10. Berman, M., and M. F. Weiss, SAAM manual, U.S. DHEW publication NIH 76-730, U.S. Government Printing Office, Washington, D.C., 1976. 11. Berman, M., Iodide kinetics, In Rail, E., and I. J. Kopin (eds.), The Thyroid and Biogenic Amines, North-Holland, Amsterdam, New York, 1972, p. 172. 12. Berman, M., The formulation and testing of models, Ann NYAcad Sci 108: 182,1963. 13. Belshaw, B. E., M. Barandes, D. V. Becker, and M. Berman, A model of iodine kinetics in the dog, Endocrinology 95: 1078, 1974. 14. Hobel, C. J., J. Sack, L. M. Cousins, and D. A. Fisher, The effect of intraamniotic thyroxine on thyroid function in the human fetus and newborn, Clin Res 25: A189, 1977. 15. Burman, K. D., J. Read, R. C. Dimond, D. Strum, F. D. Wright, W. Patow, J. M. Earll, and L. Wartofsky, Measurements of 3,3',5'-triiodothyronine (reverse TO, 3,3'-diiodothyronine, T3, and T4 in human amniotic fluid and in cord and maternal serum, J Clin Endocrinol Metab 43: 1351, 1976.


576

McGUIRE AND BERMAN

16. Chopra, I. J., and B. F. Crandall, Thyroid hormones and thyrotropin in amniotic fluid, N EnglJ Med 293: 740, 1975. 17. Fisher, D. A., I. J. Chopra, and J. H. Dussault, Extrathyroidal conversion of thyroxine to triiodothyronine in sheep, Endocrinology 91: 1141, 1972. 18. Chopra, I. J., A study of extrathyroidal conversion of thyroxine (T4) to 3,3',5-triiodothyronine (T3) in vitro, Endocrinology 101: 453, 1977. 19. Book, S. A., Transfer of radioiodine in pregnant and fetal sheep, PhD Thesis, University of California at Davis, 1973. 20. Chopra, I. J., J. Sack, and D. A. Fisher, 3,3',5'-Triiodothyronine (reverse T3) and 3,3',5-triiodothyronine (T3) in fetal and adult sheep: studies of metabolic clearance rates, production rates, serum binding, and thyroidal content relative to thyroxine, Endocrinol-

Endo • 1978 Vol 103 • No 2

ogy 97: 1080, 1975. 21. Hutchinson, D. L., M. J. Gray, A. A. Plentl, H. Alvarez, R. Caldeyro-Barcia, B. Kaplan, and J. Lind, The role of the fetus in the water exchange of the amniotic fluid of normal and hydramniotic patients, J Clin Invest 38: 971, 1959. 22. London, W. T., W. L. Money, and R. W. Rawson, Placental transfer of 131I-labeled iodide in the guineapig, J Endocrinol 28: 247, 1964. 23. Sack, J., D. A. Fisher, and R. W. Lam, Thyroid hormone metabolism in amniotic and allantoic fluids of the sheep, Pediatr Res 9: 837, 1975. 24. Miller, F. C, D. A. Sacks, S. Yeh, R. H. Paul, B. S. Schifrin, C. B. Martin, Jr., and E. H. Hon, Significance of meconium during labor, Am J Obstet Gynecol 122: 573, 1975.


Effects of dexamethasone on fetal and maternal thyroxine, triiodothyronine, reverse triiodothyronine, and thyrotropin levels.

Thyroxine uptake and metabolism by fetal sheep after intra-amniotic thyroxine injection.

Acute effects of thyroxine, triiodothyronine, and iodide on thyrotropin secretion.

Serum thyroxine and triiodothyronine concentrations during iodide treatment of hyperthyroidism.

Tetranectin in amniotic fluid, maternal serum and fetal fluids.

Effect of intraamniotic fluid thyroxine injection on fetal serum and amniotic fluid iodothyronine concentrations.

Molecular signature of amniotic fluid derived stem cells in the fetal sheep model of myelomeningocele.

Intra-amniotic thyroxine to treat fetal goiter.

Fetal-maternal fluid and electrolyte relations during chronic fetal urine loss in sheep.

Decreased reverse triiodothyronine (RT3) concentration in amniotic fluid in fetal hypothyroidism.

Standardization of amniotic fluid leptin levels and utility in maternal overweight and fetal undergrowth.

Amniotic fluid composition changes during urine drainage and tracheoesophageal occlusion in fetal sheep.

Distribution of relaxin between human maternal and fetal circulations and amniotic fluid.

[Triiodothyronine, reverse-triiodothyronine, thyroxine, resin-triiodothyronine-uptake and protein bound jodide in the fluid of thyroid cysts (author's transl)].

Catalytic ferrous iron in amniotic fluid as a predictive marker of human maternal-fetal disorders.

Maternal plasma and amniotic fluid chemokines screening in fetal Down syndrome.

3,3',5'-triiodothyronine, thyroxine, triiodothyronine, and thyrotropin levels in maternal and cord blood sera from endemic goiter regions of Brazil.

Propofol Pharmacokinetics and Estimation of Fetal Propofol Exposure during Mid-Gestational Fetal Surgery: A Maternal-Fetal Sheep Model.

Amniotic fluid catecholamine metabolites in maternal smoking.

Fetal maturity: biochemical analyses of amniotic fluid.

Decreased outer ring monodeiodination of thyroxine and reverse triiodothyronine in the fetal and neonatal rat.

Maternal administration of solithromycin, a new, potent, broad-spectrum fluoroketolide antibiotic, achieves fetal and intra-amniotic antimicrobial protection in a pregnant sheep model.

Conjugated glucocorticoids in amniotic fluid and fetal lung maturation.

Total cortisol in amniotic fluid and fetal lung maturation.