monthly cycle in the plasma concentration of GnRH and the glomerular filtered load of GnRH during
A
second trimester pregnancy Per O
Paaby, Aage Nielsen and Kjeld Raffn
Department of Clinical Chemistry, Aalborg Hospital and Department of Mathematics and Computer Science', Aalborg University, Aalborg Denmark
Paaby PO, Nielsen A, Raffn K. A monthly cycle in the plasma concentration of GnRH and the glomerular filtered load of GnRH during second trimester pregnancy. Acta Endocrinol 1992;127: 215-19. ISSN 0001-5598. Six healthy women were examined three times a week during the second trimester. The aim of the study was to test the existence of a monthly cycle in plasma GnRH. Because the kidney plays a dominant part in the elimination of small peptides from the circulation by filtration, the filtered load (P-GnRH \m=x\ GFR) of GnRH was estimated and tested for a cycle. A mathematical model (parabolas overlaid with a cosine curve and containing parameters for cycle length, cycle amplitude and phase) was used in a multivariate analysis of changes in plasma GnRH, creatinine clearance and GnRH filtered load. We found significant monthly cycles in plasma GnRH (p 0.003) and in GnRH filtered load (p 0.004), but no significance for a cycle in creatinine clearance. As an intermediate result we demonstrated that the tubular reabsorption of GnRH was unsaturated for all values of plasma GnRH and GnRH filtered load in question. The increased pregnancy level of plasma GnRH originates in part in the placenta, but we assume that the cycle generator is contained in the maternal neuroendocrine system. =
P
Paaby, Bakkevœnget 12,
9000
Aalborg,
In previous papers we demonstrated monthly cycles in creatinine clearance (1), the serum concentrations of progesterone and aldosterone and the urinary excretion rate of potassium (2), and in the serum concentration of cortisol (3) in 11 healthy women during third trimester pregnancy. The third trimester differs from the second in its considerably higher level of placental hormones, pep¬ tides as well as steroids, and it has been shown that for instance estriol undergoes an acute rise at the beginning of the third trimester with such precision that it can be used to calculate the time of parturition (4). It has also been shown that the placental response (in vitro) to the
effect of gonadotropin releasing hormone (GnRH) changes with the gestational age of the placenta (5, 6). A GnRH-like decapeptide and its prohormone is produced in the placenta (7-10), which contains recep¬ tors for that hormone (11). In vitro experiments have indicated that the production of various hormones in the placenta may be controlled by GnRH stimulation (12, 13). We therefore decided to begin our exploration of the second trimester with analysis of GnRH in plasma and urine. Creatinine clearance was also measured because the filtered load of GnRH might provide an estimate of the secretion rate of GnRH.
=
Denmark
mittee. Each
women also gave permission to use clinical information from her case record. All the women lived and worked normally without any restrictions during the study.
Protocol
Sampling of blood and urine took place three times a week during the second trimester with overlap into both the first and third trimesters. Sampling could not be completely regular because of holidays and minor intercurrent diseases. The first investigation took place as close to the beginning of the second trimester as possible, in accordance with the information in the obstetrical case record on the estimated time of parturi¬ tion. The gestational age and maturity of each offspring were controlled at the time of parturition. Urine was collected overnight in accordance with the written instructions of the Department of Clinical Chemistry and the Department of Obstetrics and Gynecology for urine collection for endogenous overnight creatinine clear¬ ance. The urine was brought to the laboratory in the morning with exact information on sampling time; the volume was measured. The number of samples in the single woman varied between 38 and 58, total number 252.
Subjects and methods Analytical methods Six healthy women, laboratory technicians and nurses, gave their informed consent to participate in the study, Creatinine clearance was carried out as overnight which was approved by the local ethics research com- creatinine clearance as described in (1), with no corree-
216
Per 0
Paaby, Aage Nielsen and Kjeld Raffn
acta ENDOCRINOLOGÍA 1992, 127
subject's height and (changing) weight and surface. GnRH was determined by radioimmunoassay (LH-RH(125I)) from Euro-Diagnostics BV, Appeldoorn, Holland. Veinpuncture was performed at an individually fixed time between 08.00 and 09.00 on the unfasting tion for the
woman.
Lithiumheparin plasma
was
separated by
immediate centrifugation at +4°C and stored at 30°C until analysis could be performed in one run on all samples from each woman. Aliquots of overnight urine were also stored deepfrozen until analysis, which was
for the single woman varied between 0.85±0.28 and 3.44±1.17 pmol/1. For all subjects P-GnRH was 1.95 ±1.23 during the last two weeks of the first trimester, 2.40 ± 1.3 9 during the first two weeks of trimester two, 1.94±0.88 during the second half of the second trimester, and 2.06 ±1.19 during the first half of the third trimester. averages
—
performed in duplicate. Intra-assay variation calculated
as
the coefficient of
variation from the differences between 39 duplicate samples from one woman with an average plasma level of GnRH of 1.8 pmol/1 was 28% (own investigation). At a
level of 20 pmol/1 the CV was 5% (own investigation) and at 200 pmol/1 4%. (Euro-Diagnostics). GnRH is unstable in biological fluids and external standards (control sera) cannot be acquired. Repeated analysis at different times on extract aliquots kept deepfrozen showed a CV of 7% at 30 pmol/1 and 5% at 150 pmol/1
Urinary excretion
rate
of GnRH
For all subjects the average urinary excretion rate of GnRH (U-GnRH) was 10.04±3.49 pmol/10 h. For the entire period the averages for the single subjects varied between 6.86±1.27 and 14.86±2.92. For all subjects the average was 8.86 ±2.17 during the last half of the first trimester, 9.62 ±2.89 during the first half of the second trimester and 10.27±3.16 during the second, and finally 10.75 ±4.54 during the first half of the third trimester.
(Euro-Diagnostics).
GnRH filtered load
Model and statistics The measurements of plasma GnRH (P-GnRH), GnRH filtered load (GnRH fl) and creatinine clearance were analysed by a multivariate model of the form
GnRH fl was calculated as P-GnRH pmol/1 x creatinine clearance ml/min x 0.6. For all subjects GnRH fl was 173.53±109.31 pmol/10 h, and for the single subjects it varied between 79.81 ±23.54 and 323.50± 150.41 pmol/10 h. In the last half of the first trimester it was 138.13±118.09, in the first and second halves of the second trimester 190.44± 136.61 and 172.28±85.25 respectively, and in the first half of the third trimester 179.20±94.68.
Eyii a.i+ßiXij + y'ixfi + okeos(2n(p¡xn + ek)) where Ey¡¡ is the expected value of the i'th persons j'th measurement of the k'th quantity. Xy is the time for the i'th persons j'th measurement, measured on a scale with 0 being the first day of the last menstruation before pregnancy and the unit being the length of each Tubular reabsorption of GnRH woman's previous menstrual cycle. Sk are amplitude parameters. There is one for each quantity, but they are GnRH reabsorption was calculated as the difference assumed to be equal for all women. The same goes for the between GnRH fl and U-GnRH, and the fractional phase parameters ek; pi are period parameters. There is reabsorption as GnRH reabsorption in % of GnRH fl. The one for each woman, but they are assumed to be equal fractional reabsorption for all subjects was for the three quantities. 90.59±9.97% ranging between 85.26±5.32 and This is the main assumption, which leads to a 96.02±2.83 in the single subjects. There was no multivariate model and makes it possible to utilize all systematic change during the gestational period investi¬ measurements and thereby arrive at a better estimate of gated in neither the single subjects or overall. the period lengths. All parameters in the model are P-GnRH and GnRH reabsorptions were related as well estimated in one run using the whole set of data. in the single subjects as in the pooled results with no the reader =
For details of the methods
appendix
is referred to the
of (2).
reabsorption
Results All deliveries were spontaneous and uncomplicated. Three boys and three girls were born. All were healthy.
Plasma concentration
of GnRH
The average P-GnRH for all subjects and all measure¬ ments (N 252) was 2.18±1.86 (SD) pmol/1. The =
indication of
a saturation of the tubular capacity for within the levels of P-GnRH and GnRH
filtered load in question (Fig.l is given as an example). Graphical comparison showed a relation between PGnRH and GnRH fl in all subjects and no relation between P-GnRH and U-GnRH or between GnRH fl and U-GnRH or between creatinine clearance and GnRH fl. In some subjects U-GnRH tended to undulate in parallel with creatinine clearance, but graphical com¬ parison between them showed a wide scatter and no relation in neither the single subjects or in the pooled results.
Monthly cycle in
acta ENDOCRINOLOGICA 1992, 127
the plasma concentration
of GnRH
217
Standard errors of the estimates. Also the p-values for tests of the amplitude parameters being zero are given. Finally, the last column in Table 1 gives the estimates of the standard deviations of the residuals. The estimates of the correlations between the residuals of pairs of quanti¬ ties are: p.GnRH/creatinine clearance 0.04, p.GnRH/ GnRHfl 0.64 and creatinine clearance/GnRHfl 0.69.
Tubular reabsorption of GnRH pmol/10h 500
4004
300H
Discussion
2001
The short intravascular half-life of GnRH and its rapid disappearance from the vascular space through filt¬ ration and hydrolysis in the kidneys makes its plasma concentration an unreliable measure of the secreted amount.
100
estimate 22.4 days. No. 3, previous 28-30, estimate 27.2. No. 4, previous 24-25, estimate 24.3. No. 5, previous 28, estimate 26.5. No. 6, previous 28-30, estimate 25. For woman no. 1 the estimated period length is about one-half the length of the previous menstrual cycle. In
The intravascular half-life of intravenously injected GnRH has been determined at 2.4±0.4 min (14) and because the estimated amounts of GnRH in plasma were low compared with the amount injected the authors suggest a rapid passage out of the vascular compart¬ ment. Others (15) found that the half-time disappear¬ ance of intravenously administered GnRH was 3.6 min and calculated the half-life due to metabolic breakdown to 4.8 min. The metabolic clearance rate (MCR) of GnRH in normal subjects was 23.7±1.8 ml-min~]-kg~' and patients with liver disease had similar values; but in patients with chronic renal failure MCR was reduced to 9.1 ±0.7 ml-min_1-kg_1, corresponding to a prolonged half disappearance time of GnRH in these patients (16). The authors conclude that the kidney is an important catabolic organ for GnRH. GnRH is a decapeptide, 1083 Dalton, and like other small linear peptides "exclusively" filtered in the glomeruli (probably freely) and reabsorbed and hydrolized in the brush border of the proximal tubuli (17). The amount found in the urine is the remnant of this process and our results show that it varies around 10% of the filtered load without any correlation between U-GnRH and P-GnRH filtered load. U-GnRH is therefore also a bad measure of the secreted amount. Under basal conditions the plasma disappearance rate of Low Molecular Weight Proteins (LMWP) is equal to the rate of secretion. The rate limiting step for removal of LMWP from the circulation is the filtration process because neither tubular absorption nor hydrolysis is saturated over a wide range of filtered loads of proteins (17). Our results show that the tubular reabsorption of GnRH was not saturated in any of the subjects, who were all healthy and had normal glomerular filtration rates as measured by creatinine clearance. We therefore
1 is excluded. We comment on this in the discussion of the statistical method. An overall multivar¬ iate test of significance for the existence of the cyclic
Table 1. Estimate of the amplitude parameter ¿for plasma GnRH, GnRH filtered load and creatinine clearance, p indicates the statistical significance of S. i. gives the estimates of the phase parameter, à gives the estimates of the standard deviation of the residuals.
0
3 P-GnRH pmol/l
12
4
5
6
Fig. 1. The relation between the plasma GnRH concentration, pmol/1 (abscissa) and the amount of GnRH absorbed, pmol/10 h (ordinate) indicates that the tubular capacity for reabsorption is unsaturated. Subject no. 2 as an example.
Overnight creatinine clearance The average was 146.77±35.21 ml/min. Creatinine clearance in the single subjects ranged between 121.78 ±11.3 5 and 176.10 ±76.44 ml/min. In the first trimester the average was 128.33±26.64, in two halves of the second trimester 147.61 ±54.71 and 151.31 ±21.81. and in the first half of the third trimester 152.34±20.10 ml/min.
Statistical results The length of the previous menstrual cycle as given by each woman compared with the cycle length estimated by the statistical method was as follows: Woman no. 1, previous cycle irregular 28-36 days, estimate 13.7 days. No. 2, previous cycle always irregular 23-33 days,
the statistical
analysis supporting the following results,
woman no.
gives a highly significant result (p 0.0005). In Fig. 2 the concentrations of P-GnRH and the cyclic curve estimated by the mathematical
components
(5±SF,
=
model are shown for woman no. 2. Table 1 gives the estimated values of the amplitude parameters and the phase parameters together with the
P-GnRH
GnRHfl Creatinine clearance
p
£±SE
(T
0.24±0.07 0.003 0.230±0.()48 0.72 33.30±9.80 0.004 0.230±0.050 99.90 2.48±3.23 0.750 0.038±0.200 31.70
218
Per O
Paaby, Aage Nielsen and Kjeld Raffn
acta endocrinoloüica 1992. 127
H 0
9r-.
Fig. 2. The figure illustrates woman no. 2. Plasma concentrations of GnRH (black dots connected with broken lines) and general trend with the cyclic component. Note that the cyclic component is estimated from the data of all five women (see text). Woman no. 2 had an irregular previous menstrual cycle of 23-3 3 days. The not shown zero on the abscissa is the first day of the last menstruation before pregnancy, and the unit is 28 days. The cycle length estimated from the data is 22.4 days. Ordinate plasma GnRH. pmol/1.
assume
that the filtered load of GnRH
measures
the
disappearance rate and hence the secretion rate of GnRH and we believe that our finding of significant monthly cycles in the GnRH filtered load indicates a monthly cycle in the average level of secretion of GnRH. We do not suggest that the secretion runs along a
smooth sinosoid curve. It is well known that GnRH is secreted pulsatively with from 1 to about 3 h between pulses (18). The irregular pattern of our P-GnRH curves clearly indicates the existence of pulses with major surges in between (Fig. 2). Sampling three times a week does not of course allow a detailed description of the pulsative pattern, but a more frequent sampling over a long period of pregnancy is not permissible. We must accept the pulsative secretion as noise on the statistical analysis together with the analytical variation. The high levels of significance, in spite of the noise, support our belief that the rate of GnRH- secretion runs along a higher level at regular monthly intervals. This may be difficult to accept because of the conflicting evidence of a monthly menstrual cycle in GnRH. GnRH secretion is often gauged by changes in LH. Experiments on ewes have shown that LH pulses in peripheral blood are always preceded by a GnRH pulse (19). The highest frequency and amplitude of LH pulses during the follicular phase of the human menstrual cycle occur in the late follicular phase. During the luteal phase the frequency is lower but the highest amplitude and frequency appear in the early luteal phase (18). Estro¬ gens inhibit the human GnRH pulse generator (20), and during the menstrual cycle GnRH pulses occur with a lower frequency but with a higher amplitude during the luteal phase (21).
In the present study GnRH was the only hormone studied, and we therefore cannot know how the GnRH
cycle is "phased" in relation to other hormones. We intend to explore this in our next study, and we suggest that a possible study of GnRH secretion during the menstrual cycle should include GnRH fl. Woman nos. 1 and 2 had markedly irregular previous menstrual cycles. In no. 1 the estimated cycle length was about half the minimum length of the previous cycle and in no. 2 it was equal to the minimum of the previous cycle. None of the subjects described in our previous studies (1-3) had irregular menstrual cycles, and the estimated cycle lengths agreed with the lengths of the previous cycles as in woman nos. 3, 4 5 and 6 in the present study. The finding of significant hormonal cycles during pregnancy indicates that the "pregnancy cycle" is independent of the ovarian oscillator. We cannot know if this is of greater consequence especially in women with ,
irregular cycles. In this study we did not find evidence of a monthly cycle in creatinine clearance. We cannot explain this; we can only point to the fact that the level of significance for a cycle in creatinine clearance found in (1) was much lower than the significance for hormone cycles found in (2) or (3), and also refer to a much lower number of subjects than in (2, 3).
In calculation of GnRH fl, creatinine clearance was used as an estimate of GFR. Repeated use of tracer methods for the estimation of "true" GFR cannot be used three times a week during pregnancy, and we have previously shown that creatinine clearance changes like 51Cr-EDTA clearance during the menstrual cycle (22).
Monthly cycle in
ACTA ENDOCRINOLOGÍA 1992, 127
Discussion
of the statistical method
to validate the statistical procedure performed mainly as those described in the appen¬ dix of (2). No serious departures from the assumptions underlying the statistical model were found. Although the amplitudes for P-GnRH and GnRH fl are highly significant the model leaves a lot of variation unexplained. This is seen from the fairly large values
A
variety of checks
were
of the estimated standard deviations of the residuals (Table 1). Out ofthe variation about the parabolas which model the general trend, the cyclic components explain approximately 4% for P-GnRH and GnRH fl and less than 1% for creatinine clearance. This should be kept in mind when looking at Fig. 2. We do not claim that the cyclic components are easily detectable by eye, but we are sure
they are present. We paid special attention to woman no. 1 whose estimated period length was about one-half the length of the previous menstrual cycle. From our visual examin¬ ation of previous (1-3) and present data we got the impression that such "half period undulations" existed. We do think that woman no. 1 has a cyclic variation with a period close to the previous menstrual cycle, but the iterative estimation procedure has "caught" the half period component for this woman. This is confirmed by the fact that if we perform the overall test of significance for the existence of the cyclic components including woman no. 1 imposing on her twice the estimated period length, then we get an even more significant result (p 0.0002) than when she is excluded. We conclude that significant monthly cycles in plasma GnRH (p 0.003) and in GnRH filtered load (p 0.004) exist. As an intermediate result we demon¬ strated that the tubular reabsorption of GnRH was unsaturated for all values of plasma GnRH and GnRH =
=
=
filtered load in question. The increased pregnancy level of plasma GnRH originates in part in the placenta, but we suppose that the cycle generator is contained in the maternal neuroendocrine system.
Acknowledgments. This study was supported by grants from the Novo Foundation, Danish Foundation for the Advancement of Medical Science (AP Möller), the Foundation of Volunteer Blood Donors, Aalborg, the Memorial Foundation of JM Klein and wife, and the Foundation of Inga and Vagn Larsen, Aalborg.
References 1.
Paaby P, Nielsen Aa, Moller-Petersen J, Raffn K. Cyclical changes in endogenous overnight creatinine clearance during the third
2.
Paaby
3.
the
219
4. Buster JE. Freeman AG et al. Time trend analysis of unconjugated estriol concentrations in third trimester pregnancy. Obstet Gynecol 1980:56:743-7 5. Siler-Khodr TM, Khodr GS, Valenzuela G, Harper MJK. Rhode J. GnRH effects on placental hormones during gestation. III. Biol 6.
7.
Reprod 1986:35:312-19 Haning RV, Breault PH, Mahendra V, DeSilva, Hackett RJ, Pouncey CL. Effects of fetal sex, stage of gestation dibutyryl cyclic adenosine monophosphate and gonadotropin releasing hormone on secretion of human chorinoic gonadotropin by placental expiants in vitro. Am J Obstet Gynecol 1988:159:1 332-7 Khodr GS, Siler-Khodr TM. Localization of luteinizing hormonereleasing factor in the human placenta. Fértil Steril 1978:29:523-6
Hypothalamic-like releasing hormones of the placenta. Clin Perinatal 1983:10:553-66 Tan L, Rousseau P. The chemical identity of the immuno-reactive LHRH-like peptide biosynthesized in human placenta. Biochem Biophys Res Commun 1982;109:1061-71 Seeburg PH, Adelman JP. Characterization of cDNA for precursor of human luteinizing hormone releasing hormone. Nature 1984;
8. Siler-Khodr TM. 9.
10.
311:666-8 11. Iwashita M, Evans MI, Catt KJ. Characterization of gonadotropinreleasing hormone receptor site in term placenta and chorionic villi. J Clin Endocrinol Metab 1986:62:127-33 12. Siler-Khodr CM. Khodr GS, Valenzuela G, Rhode J. Gonadotropinreleasing hormone effects on placental hormones during ges¬ tation. I. Biol Reprod 1986;34:245-54 1 3. Siler-Khodr CM. Khodr GS, Valenzuela G, Rhode J. Gonadotropinreleasing hormone effects on placental hormones during ges¬ tation. II. Biol Reprod 1986;34:255-64 14. Fauconnier JP, Teuwissen B, Thomas K. Rate of disappearance in plasma of synthetic LH-RH intravenously injected in man. Gynecol Obstet Invest 1978:9:229-37 15. Arimura A, Kastin AJ, González-Barcena D, Siller J, Weaver RE, Schally AV. Disappearance of LH-releasing hormone in man as determined by radioimmunoassay. Clin Endocrinol 1974; 3: 421-5 16. Pimstone B, Epstein S, Hamilton D, LeRoith D, Hendricks S. Metabolic clearance and plasma half disapperance time of exoge¬ nous gonadotropin releasing hormone in normal subjects and in patients with liver disease and chronic renal failure. J Clin Endocrinol Metab 1977:44:356-60 17. The kidney. Physiology and pathophysiology. New York: Raven Press 1985:1796-9 18. Filiacori M. Flamigni C, Vizziello C et al. Hypothalamic control of gonadotropin secretion in the human menstrual cycle. Prog Clin Biol Res 1986:225:55-74 19. Levine JE, Pau KF, Ramirez UD, Jackson GL. Simultaneous measurement of luteinizing hormone release in the unanesthetized ovariectomised sheep. Endocrinology 1982;111:1449-55 20. Veldhuis JD. The hypothalamic pulse generator: the reproductive core. Clin Obstet Gynecol 1990:33:538-50 21. Jaffe RB, Plosker S, Marshall L, Martin MC. Neuromodulatory regulation of gonadotropin-releasing hormone pulsatile discharge in women. Am J Obstet Gynecol 1990:163:1 727-31 22. Paaby P, Brochner-Mortensen J, Fjeldborg P, Raffn K, Larsen CE, Moller-Petersen J. Endogenous overnight creatinine clearance compared with 51Cr-Edta clearance during the menstrual cycle. Acta Med Scand 1987:222:281-4
trimester of pregnancy. Acta Med Scand 1988:223:459-68 P, Nielsen Aa, Raffn K. A monthly cycle during third trimester pregnancy in the serum concentrations of progesterone and aldosterone and the urinary excretion rate of potassium. Acta
Endocrinol (Copenh) 1989;120:636-43 Paaby P, Nielsen Aa, Raffn K. A monthly cycle in the adrenocortical function during third trimester pregnancy. Acta Endocrinol (Copenh) 1990;122:617-22
plasma concentration of GnRH
January 20th, 1992 Accepted April 28th, 1992 Received
A
second trimester pregnancy Per O
Paaby, Aage Nielsen and Kjeld Raffn
Department of Clinical Chemistry, Aalborg Hospital and Department of Mathematics and Computer Science', Aalborg University, Aalborg Denmark
Paaby PO, Nielsen A, Raffn K. A monthly cycle in the plasma concentration of GnRH and the glomerular filtered load of GnRH during second trimester pregnancy. Acta Endocrinol 1992;127: 215-19. ISSN 0001-5598. Six healthy women were examined three times a week during the second trimester. The aim of the study was to test the existence of a monthly cycle in plasma GnRH. Because the kidney plays a dominant part in the elimination of small peptides from the circulation by filtration, the filtered load (P-GnRH \m=x\ GFR) of GnRH was estimated and tested for a cycle. A mathematical model (parabolas overlaid with a cosine curve and containing parameters for cycle length, cycle amplitude and phase) was used in a multivariate analysis of changes in plasma GnRH, creatinine clearance and GnRH filtered load. We found significant monthly cycles in plasma GnRH (p 0.003) and in GnRH filtered load (p 0.004), but no significance for a cycle in creatinine clearance. As an intermediate result we demonstrated that the tubular reabsorption of GnRH was unsaturated for all values of plasma GnRH and GnRH filtered load in question. The increased pregnancy level of plasma GnRH originates in part in the placenta, but we assume that the cycle generator is contained in the maternal neuroendocrine system. =
P
Paaby, Bakkevœnget 12,
9000
Aalborg,
In previous papers we demonstrated monthly cycles in creatinine clearance (1), the serum concentrations of progesterone and aldosterone and the urinary excretion rate of potassium (2), and in the serum concentration of cortisol (3) in 11 healthy women during third trimester pregnancy. The third trimester differs from the second in its considerably higher level of placental hormones, pep¬ tides as well as steroids, and it has been shown that for instance estriol undergoes an acute rise at the beginning of the third trimester with such precision that it can be used to calculate the time of parturition (4). It has also been shown that the placental response (in vitro) to the
effect of gonadotropin releasing hormone (GnRH) changes with the gestational age of the placenta (5, 6). A GnRH-like decapeptide and its prohormone is produced in the placenta (7-10), which contains recep¬ tors for that hormone (11). In vitro experiments have indicated that the production of various hormones in the placenta may be controlled by GnRH stimulation (12, 13). We therefore decided to begin our exploration of the second trimester with analysis of GnRH in plasma and urine. Creatinine clearance was also measured because the filtered load of GnRH might provide an estimate of the secretion rate of GnRH.
=
Denmark
mittee. Each
women also gave permission to use clinical information from her case record. All the women lived and worked normally without any restrictions during the study.
Protocol
Sampling of blood and urine took place three times a week during the second trimester with overlap into both the first and third trimesters. Sampling could not be completely regular because of holidays and minor intercurrent diseases. The first investigation took place as close to the beginning of the second trimester as possible, in accordance with the information in the obstetrical case record on the estimated time of parturi¬ tion. The gestational age and maturity of each offspring were controlled at the time of parturition. Urine was collected overnight in accordance with the written instructions of the Department of Clinical Chemistry and the Department of Obstetrics and Gynecology for urine collection for endogenous overnight creatinine clear¬ ance. The urine was brought to the laboratory in the morning with exact information on sampling time; the volume was measured. The number of samples in the single woman varied between 38 and 58, total number 252.
Subjects and methods Analytical methods Six healthy women, laboratory technicians and nurses, gave their informed consent to participate in the study, Creatinine clearance was carried out as overnight which was approved by the local ethics research com- creatinine clearance as described in (1), with no corree-
216
Per 0
Paaby, Aage Nielsen and Kjeld Raffn
acta ENDOCRINOLOGÍA 1992, 127
subject's height and (changing) weight and surface. GnRH was determined by radioimmunoassay (LH-RH(125I)) from Euro-Diagnostics BV, Appeldoorn, Holland. Veinpuncture was performed at an individually fixed time between 08.00 and 09.00 on the unfasting tion for the
woman.
Lithiumheparin plasma
was
separated by
immediate centrifugation at +4°C and stored at 30°C until analysis could be performed in one run on all samples from each woman. Aliquots of overnight urine were also stored deepfrozen until analysis, which was
for the single woman varied between 0.85±0.28 and 3.44±1.17 pmol/1. For all subjects P-GnRH was 1.95 ±1.23 during the last two weeks of the first trimester, 2.40 ± 1.3 9 during the first two weeks of trimester two, 1.94±0.88 during the second half of the second trimester, and 2.06 ±1.19 during the first half of the third trimester. averages
—
performed in duplicate. Intra-assay variation calculated
as
the coefficient of
variation from the differences between 39 duplicate samples from one woman with an average plasma level of GnRH of 1.8 pmol/1 was 28% (own investigation). At a
level of 20 pmol/1 the CV was 5% (own investigation) and at 200 pmol/1 4%. (Euro-Diagnostics). GnRH is unstable in biological fluids and external standards (control sera) cannot be acquired. Repeated analysis at different times on extract aliquots kept deepfrozen showed a CV of 7% at 30 pmol/1 and 5% at 150 pmol/1
Urinary excretion
rate
of GnRH
For all subjects the average urinary excretion rate of GnRH (U-GnRH) was 10.04±3.49 pmol/10 h. For the entire period the averages for the single subjects varied between 6.86±1.27 and 14.86±2.92. For all subjects the average was 8.86 ±2.17 during the last half of the first trimester, 9.62 ±2.89 during the first half of the second trimester and 10.27±3.16 during the second, and finally 10.75 ±4.54 during the first half of the third trimester.
(Euro-Diagnostics).
GnRH filtered load
Model and statistics The measurements of plasma GnRH (P-GnRH), GnRH filtered load (GnRH fl) and creatinine clearance were analysed by a multivariate model of the form
GnRH fl was calculated as P-GnRH pmol/1 x creatinine clearance ml/min x 0.6. For all subjects GnRH fl was 173.53±109.31 pmol/10 h, and for the single subjects it varied between 79.81 ±23.54 and 323.50± 150.41 pmol/10 h. In the last half of the first trimester it was 138.13±118.09, in the first and second halves of the second trimester 190.44± 136.61 and 172.28±85.25 respectively, and in the first half of the third trimester 179.20±94.68.
Eyii a.i+ßiXij + y'ixfi + okeos(2n(p¡xn + ek)) where Ey¡¡ is the expected value of the i'th persons j'th measurement of the k'th quantity. Xy is the time for the i'th persons j'th measurement, measured on a scale with 0 being the first day of the last menstruation before pregnancy and the unit being the length of each Tubular reabsorption of GnRH woman's previous menstrual cycle. Sk are amplitude parameters. There is one for each quantity, but they are GnRH reabsorption was calculated as the difference assumed to be equal for all women. The same goes for the between GnRH fl and U-GnRH, and the fractional phase parameters ek; pi are period parameters. There is reabsorption as GnRH reabsorption in % of GnRH fl. The one for each woman, but they are assumed to be equal fractional reabsorption for all subjects was for the three quantities. 90.59±9.97% ranging between 85.26±5.32 and This is the main assumption, which leads to a 96.02±2.83 in the single subjects. There was no multivariate model and makes it possible to utilize all systematic change during the gestational period investi¬ measurements and thereby arrive at a better estimate of gated in neither the single subjects or overall. the period lengths. All parameters in the model are P-GnRH and GnRH reabsorptions were related as well estimated in one run using the whole set of data. in the single subjects as in the pooled results with no the reader =
For details of the methods
appendix
is referred to the
of (2).
reabsorption
Results All deliveries were spontaneous and uncomplicated. Three boys and three girls were born. All were healthy.
Plasma concentration
of GnRH
The average P-GnRH for all subjects and all measure¬ ments (N 252) was 2.18±1.86 (SD) pmol/1. The =
indication of
a saturation of the tubular capacity for within the levels of P-GnRH and GnRH
filtered load in question (Fig.l is given as an example). Graphical comparison showed a relation between PGnRH and GnRH fl in all subjects and no relation between P-GnRH and U-GnRH or between GnRH fl and U-GnRH or between creatinine clearance and GnRH fl. In some subjects U-GnRH tended to undulate in parallel with creatinine clearance, but graphical com¬ parison between them showed a wide scatter and no relation in neither the single subjects or in the pooled results.
Monthly cycle in
acta ENDOCRINOLOGICA 1992, 127
the plasma concentration
of GnRH
217
Standard errors of the estimates. Also the p-values for tests of the amplitude parameters being zero are given. Finally, the last column in Table 1 gives the estimates of the standard deviations of the residuals. The estimates of the correlations between the residuals of pairs of quanti¬ ties are: p.GnRH/creatinine clearance 0.04, p.GnRH/ GnRHfl 0.64 and creatinine clearance/GnRHfl 0.69.
Tubular reabsorption of GnRH pmol/10h 500
4004
300H
Discussion
2001
The short intravascular half-life of GnRH and its rapid disappearance from the vascular space through filt¬ ration and hydrolysis in the kidneys makes its plasma concentration an unreliable measure of the secreted amount.
100
estimate 22.4 days. No. 3, previous 28-30, estimate 27.2. No. 4, previous 24-25, estimate 24.3. No. 5, previous 28, estimate 26.5. No. 6, previous 28-30, estimate 25. For woman no. 1 the estimated period length is about one-half the length of the previous menstrual cycle. In
The intravascular half-life of intravenously injected GnRH has been determined at 2.4±0.4 min (14) and because the estimated amounts of GnRH in plasma were low compared with the amount injected the authors suggest a rapid passage out of the vascular compart¬ ment. Others (15) found that the half-time disappear¬ ance of intravenously administered GnRH was 3.6 min and calculated the half-life due to metabolic breakdown to 4.8 min. The metabolic clearance rate (MCR) of GnRH in normal subjects was 23.7±1.8 ml-min~]-kg~' and patients with liver disease had similar values; but in patients with chronic renal failure MCR was reduced to 9.1 ±0.7 ml-min_1-kg_1, corresponding to a prolonged half disappearance time of GnRH in these patients (16). The authors conclude that the kidney is an important catabolic organ for GnRH. GnRH is a decapeptide, 1083 Dalton, and like other small linear peptides "exclusively" filtered in the glomeruli (probably freely) and reabsorbed and hydrolized in the brush border of the proximal tubuli (17). The amount found in the urine is the remnant of this process and our results show that it varies around 10% of the filtered load without any correlation between U-GnRH and P-GnRH filtered load. U-GnRH is therefore also a bad measure of the secreted amount. Under basal conditions the plasma disappearance rate of Low Molecular Weight Proteins (LMWP) is equal to the rate of secretion. The rate limiting step for removal of LMWP from the circulation is the filtration process because neither tubular absorption nor hydrolysis is saturated over a wide range of filtered loads of proteins (17). Our results show that the tubular reabsorption of GnRH was not saturated in any of the subjects, who were all healthy and had normal glomerular filtration rates as measured by creatinine clearance. We therefore
1 is excluded. We comment on this in the discussion of the statistical method. An overall multivar¬ iate test of significance for the existence of the cyclic
Table 1. Estimate of the amplitude parameter ¿for plasma GnRH, GnRH filtered load and creatinine clearance, p indicates the statistical significance of S. i. gives the estimates of the phase parameter, à gives the estimates of the standard deviation of the residuals.
0
3 P-GnRH pmol/l
12
4
5
6
Fig. 1. The relation between the plasma GnRH concentration, pmol/1 (abscissa) and the amount of GnRH absorbed, pmol/10 h (ordinate) indicates that the tubular capacity for reabsorption is unsaturated. Subject no. 2 as an example.
Overnight creatinine clearance The average was 146.77±35.21 ml/min. Creatinine clearance in the single subjects ranged between 121.78 ±11.3 5 and 176.10 ±76.44 ml/min. In the first trimester the average was 128.33±26.64, in two halves of the second trimester 147.61 ±54.71 and 151.31 ±21.81. and in the first half of the third trimester 152.34±20.10 ml/min.
Statistical results The length of the previous menstrual cycle as given by each woman compared with the cycle length estimated by the statistical method was as follows: Woman no. 1, previous cycle irregular 28-36 days, estimate 13.7 days. No. 2, previous cycle always irregular 23-33 days,
the statistical
analysis supporting the following results,
woman no.
gives a highly significant result (p 0.0005). In Fig. 2 the concentrations of P-GnRH and the cyclic curve estimated by the mathematical
components
(5±SF,
=
model are shown for woman no. 2. Table 1 gives the estimated values of the amplitude parameters and the phase parameters together with the
P-GnRH
GnRHfl Creatinine clearance
p
£±SE
(T
0.24±0.07 0.003 0.230±0.()48 0.72 33.30±9.80 0.004 0.230±0.050 99.90 2.48±3.23 0.750 0.038±0.200 31.70
218
Per O
Paaby, Aage Nielsen and Kjeld Raffn
acta endocrinoloüica 1992. 127
H 0
9r-.
Fig. 2. The figure illustrates woman no. 2. Plasma concentrations of GnRH (black dots connected with broken lines) and general trend with the cyclic component. Note that the cyclic component is estimated from the data of all five women (see text). Woman no. 2 had an irregular previous menstrual cycle of 23-3 3 days. The not shown zero on the abscissa is the first day of the last menstruation before pregnancy, and the unit is 28 days. The cycle length estimated from the data is 22.4 days. Ordinate plasma GnRH. pmol/1.
assume
that the filtered load of GnRH
measures
the
disappearance rate and hence the secretion rate of GnRH and we believe that our finding of significant monthly cycles in the GnRH filtered load indicates a monthly cycle in the average level of secretion of GnRH. We do not suggest that the secretion runs along a
smooth sinosoid curve. It is well known that GnRH is secreted pulsatively with from 1 to about 3 h between pulses (18). The irregular pattern of our P-GnRH curves clearly indicates the existence of pulses with major surges in between (Fig. 2). Sampling three times a week does not of course allow a detailed description of the pulsative pattern, but a more frequent sampling over a long period of pregnancy is not permissible. We must accept the pulsative secretion as noise on the statistical analysis together with the analytical variation. The high levels of significance, in spite of the noise, support our belief that the rate of GnRH- secretion runs along a higher level at regular monthly intervals. This may be difficult to accept because of the conflicting evidence of a monthly menstrual cycle in GnRH. GnRH secretion is often gauged by changes in LH. Experiments on ewes have shown that LH pulses in peripheral blood are always preceded by a GnRH pulse (19). The highest frequency and amplitude of LH pulses during the follicular phase of the human menstrual cycle occur in the late follicular phase. During the luteal phase the frequency is lower but the highest amplitude and frequency appear in the early luteal phase (18). Estro¬ gens inhibit the human GnRH pulse generator (20), and during the menstrual cycle GnRH pulses occur with a lower frequency but with a higher amplitude during the luteal phase (21).
In the present study GnRH was the only hormone studied, and we therefore cannot know how the GnRH
cycle is "phased" in relation to other hormones. We intend to explore this in our next study, and we suggest that a possible study of GnRH secretion during the menstrual cycle should include GnRH fl. Woman nos. 1 and 2 had markedly irregular previous menstrual cycles. In no. 1 the estimated cycle length was about half the minimum length of the previous cycle and in no. 2 it was equal to the minimum of the previous cycle. None of the subjects described in our previous studies (1-3) had irregular menstrual cycles, and the estimated cycle lengths agreed with the lengths of the previous cycles as in woman nos. 3, 4 5 and 6 in the present study. The finding of significant hormonal cycles during pregnancy indicates that the "pregnancy cycle" is independent of the ovarian oscillator. We cannot know if this is of greater consequence especially in women with ,
irregular cycles. In this study we did not find evidence of a monthly cycle in creatinine clearance. We cannot explain this; we can only point to the fact that the level of significance for a cycle in creatinine clearance found in (1) was much lower than the significance for hormone cycles found in (2) or (3), and also refer to a much lower number of subjects than in (2, 3).
In calculation of GnRH fl, creatinine clearance was used as an estimate of GFR. Repeated use of tracer methods for the estimation of "true" GFR cannot be used three times a week during pregnancy, and we have previously shown that creatinine clearance changes like 51Cr-EDTA clearance during the menstrual cycle (22).
Monthly cycle in
ACTA ENDOCRINOLOGÍA 1992, 127
Discussion
of the statistical method
to validate the statistical procedure performed mainly as those described in the appen¬ dix of (2). No serious departures from the assumptions underlying the statistical model were found. Although the amplitudes for P-GnRH and GnRH fl are highly significant the model leaves a lot of variation unexplained. This is seen from the fairly large values
A
variety of checks
were
of the estimated standard deviations of the residuals (Table 1). Out ofthe variation about the parabolas which model the general trend, the cyclic components explain approximately 4% for P-GnRH and GnRH fl and less than 1% for creatinine clearance. This should be kept in mind when looking at Fig. 2. We do not claim that the cyclic components are easily detectable by eye, but we are sure
they are present. We paid special attention to woman no. 1 whose estimated period length was about one-half the length of the previous menstrual cycle. From our visual examin¬ ation of previous (1-3) and present data we got the impression that such "half period undulations" existed. We do think that woman no. 1 has a cyclic variation with a period close to the previous menstrual cycle, but the iterative estimation procedure has "caught" the half period component for this woman. This is confirmed by the fact that if we perform the overall test of significance for the existence of the cyclic components including woman no. 1 imposing on her twice the estimated period length, then we get an even more significant result (p 0.0002) than when she is excluded. We conclude that significant monthly cycles in plasma GnRH (p 0.003) and in GnRH filtered load (p 0.004) exist. As an intermediate result we demon¬ strated that the tubular reabsorption of GnRH was unsaturated for all values of plasma GnRH and GnRH =
=
=
filtered load in question. The increased pregnancy level of plasma GnRH originates in part in the placenta, but we suppose that the cycle generator is contained in the maternal neuroendocrine system.
Acknowledgments. This study was supported by grants from the Novo Foundation, Danish Foundation for the Advancement of Medical Science (AP Möller), the Foundation of Volunteer Blood Donors, Aalborg, the Memorial Foundation of JM Klein and wife, and the Foundation of Inga and Vagn Larsen, Aalborg.
References 1.
Paaby P, Nielsen Aa, Moller-Petersen J, Raffn K. Cyclical changes in endogenous overnight creatinine clearance during the third
2.
Paaby
3.
the
219
4. Buster JE. Freeman AG et al. Time trend analysis of unconjugated estriol concentrations in third trimester pregnancy. Obstet Gynecol 1980:56:743-7 5. Siler-Khodr TM, Khodr GS, Valenzuela G, Harper MJK. Rhode J. GnRH effects on placental hormones during gestation. III. Biol 6.
7.
Reprod 1986:35:312-19 Haning RV, Breault PH, Mahendra V, DeSilva, Hackett RJ, Pouncey CL. Effects of fetal sex, stage of gestation dibutyryl cyclic adenosine monophosphate and gonadotropin releasing hormone on secretion of human chorinoic gonadotropin by placental expiants in vitro. Am J Obstet Gynecol 1988:159:1 332-7 Khodr GS, Siler-Khodr TM. Localization of luteinizing hormonereleasing factor in the human placenta. Fértil Steril 1978:29:523-6
Hypothalamic-like releasing hormones of the placenta. Clin Perinatal 1983:10:553-66 Tan L, Rousseau P. The chemical identity of the immuno-reactive LHRH-like peptide biosynthesized in human placenta. Biochem Biophys Res Commun 1982;109:1061-71 Seeburg PH, Adelman JP. Characterization of cDNA for precursor of human luteinizing hormone releasing hormone. Nature 1984;
8. Siler-Khodr TM. 9.
10.
311:666-8 11. Iwashita M, Evans MI, Catt KJ. Characterization of gonadotropinreleasing hormone receptor site in term placenta and chorionic villi. J Clin Endocrinol Metab 1986:62:127-33 12. Siler-Khodr CM. Khodr GS, Valenzuela G, Rhode J. Gonadotropinreleasing hormone effects on placental hormones during ges¬ tation. I. Biol Reprod 1986;34:245-54 1 3. Siler-Khodr CM. Khodr GS, Valenzuela G, Rhode J. Gonadotropinreleasing hormone effects on placental hormones during ges¬ tation. II. Biol Reprod 1986;34:255-64 14. Fauconnier JP, Teuwissen B, Thomas K. Rate of disappearance in plasma of synthetic LH-RH intravenously injected in man. Gynecol Obstet Invest 1978:9:229-37 15. Arimura A, Kastin AJ, González-Barcena D, Siller J, Weaver RE, Schally AV. Disappearance of LH-releasing hormone in man as determined by radioimmunoassay. Clin Endocrinol 1974; 3: 421-5 16. Pimstone B, Epstein S, Hamilton D, LeRoith D, Hendricks S. Metabolic clearance and plasma half disapperance time of exoge¬ nous gonadotropin releasing hormone in normal subjects and in patients with liver disease and chronic renal failure. J Clin Endocrinol Metab 1977:44:356-60 17. The kidney. Physiology and pathophysiology. New York: Raven Press 1985:1796-9 18. Filiacori M. Flamigni C, Vizziello C et al. Hypothalamic control of gonadotropin secretion in the human menstrual cycle. Prog Clin Biol Res 1986:225:55-74 19. Levine JE, Pau KF, Ramirez UD, Jackson GL. Simultaneous measurement of luteinizing hormone release in the unanesthetized ovariectomised sheep. Endocrinology 1982;111:1449-55 20. Veldhuis JD. The hypothalamic pulse generator: the reproductive core. Clin Obstet Gynecol 1990:33:538-50 21. Jaffe RB, Plosker S, Marshall L, Martin MC. Neuromodulatory regulation of gonadotropin-releasing hormone pulsatile discharge in women. Am J Obstet Gynecol 1990:163:1 727-31 22. Paaby P, Brochner-Mortensen J, Fjeldborg P, Raffn K, Larsen CE, Moller-Petersen J. Endogenous overnight creatinine clearance compared with 51Cr-Edta clearance during the menstrual cycle. Acta Med Scand 1987:222:281-4
trimester of pregnancy. Acta Med Scand 1988:223:459-68 P, Nielsen Aa, Raffn K. A monthly cycle during third trimester pregnancy in the serum concentrations of progesterone and aldosterone and the urinary excretion rate of potassium. Acta
Endocrinol (Copenh) 1989;120:636-43 Paaby P, Nielsen Aa, Raffn K. A monthly cycle in the adrenocortical function during third trimester pregnancy. Acta Endocrinol (Copenh) 1990;122:617-22
plasma concentration of GnRH
January 20th, 1992 Accepted April 28th, 1992 Received
