- Email: [email protected]
Potential abnormalities in molecular forms of growth hormone in schizophrenia: A pilot study
BIOL PSYCHIATRY 1993;34:487-491
4.87
Potential Abnormalities in Molecular Forms of Growth Hormone in Schizophrenia: A Pilot Study M. Dhyanne Warner, Yagya N. Sinha, and Cecilia A. Peabody
Growth hormone has been investigated in numerous studies involving patients with schizophrenia but has been measured only by radioimmunoassay (RIA). There have been no consistent abnormalities differentiating patients with schizophrenia from normal controls. In the current study, growth hormone (GH ) variants were measured by Western blotting techniques, which resulted in the quantitation of 4 GH size variants: 27K (27,000 Daltons), 22K, 2OK, and i 7K. in the entire sample of ! 7 schizophrenic subjects, all GH variants were significantly higher than in the 14 normal controls. While there were no significant differcnces between the 2 groups in RIA GH values, the RIA values were generally higher in the schizophrenic group. In a subset of 12 schizophrenic patients whose RIA values were approximately equal to the controls, both the 271( and 22K GH variants remained significantly higher in the patient group. In the schizophrenic group, none of the GH variants or RIA GH changed significantly after I week of treatment with neuroleptic medication. These preliminary results suggest that certain GH forms may be elevated in schizophrenia, but further studies are needed. Key Words: Schizophrenia, growth hormone Introduction Numerous studies have investigated growth hormone (GH) in schizophrenia, in part because a dopamine dysregulation theory remains a major pathophysiologic component in schizophrenia, and dopamine stimulates GH. To date there have been no consistent abnormalities to differentiate patients with schizophrenia from normal controls in either basal or stimulated conditions. GH levels have been measured by radioimmunoassay (RIA) only, with the exception of our own previous work (Peabody et al 1990). There have been several molecular forms (variants) of immunoreactive GH identified, with marked differences in molecular weight and biological and immunological activities
From the D e ~ e n t of Psychiatry, StanfordMedical School, Stanford, CA (MDW); The Whittier Institute, La Jolla, CA (YNS); and Department of Fsychiatry, VA Medical Center, Menlo Park, CA (CAP). Address correspondence to Cecilia Peabody MD, V.A, Medical Center, 324-EI03, 795 Willow Road, Menlo Park, CA 94025. Received June 22, 1992; revised June 24, 1993. © 1993 Society of Biological Psychiatry
(Baumatln 1991). Hence, some variants of GH may be abnormal but not detected by standard RIA techniques. Further definition of GH variants may provide some clues to understanding the inconsistent RIA GH results in schizophrenia. In the present study GH was measured by RIA as well as by Western blotting techniques, which resulted in 4 GH size variants identified by molecular weight: 27K, 22K, 20K and 17K. These were measured in normal controls and drug free schizophrenic patients before and 1 week after treatment with neuroleptic medication. Subjects and Methods Subjects Seventeen inpatient men with a DSM-III-R diagnosis of chronic schizophrenic participated in this study. All were at least 2 weeks free from oral and 3 months free from decanoate neuroleptics. The mean age of the schizophrenic 0006-3223/93/$06.00
488
BIOLPSYCHIATRY
M. Dhyanne Warner et al
1993;34:487-491
group was 34 years (range ! 8-47). All subjects were nonobese (< 130% ideal weight), in excellent health, and followed unrestricted dietary regimens. Five patients were drug naive, and 12 patients had been free of neuroleptics for an average of 1.6 years. The average age of onset was 25 years, and the average duration of illness was 8 years. Fourteen normal men served as controls, were in excellent health by history, were on no medication, and had no history of psychiatric illness. The mean age was 32 years (range 23-52). All subjects gave written informed consent after the protocol had been fully explained.
1
Statistics The Mann Whitney U-test was used to determine differences between the controls and schizophrenics since the groups were not normally distributed. A paired t-test was used to determine the differences between the pre- and posttreatment groups since the differences here were. essentially normally distributed. Because of multiple t-tests (5) in this analysis the conservative Bonferroni correction was used and all p values were multiplied by 5.
Results In the entire sample of 17 schizophrenic subjects, all GH variants were significantly higher than in the 14 normal controls but no differences were found for the RIA GH.
4
-27K GH
All inpatient serum samples were drawn at approximately 7:30 a.m. The blood for the normal controls was drawn throughout the day. The patient's blood was again drawn I week after starting medication. Neuroleptics were started after the baseline serum levels were drawn and doses were determined clinically. Thirteen subjects were on haloperidol (Haldol) in doses up to 30 mg per day. In 4 patients haloperidol was not deemed clinically appropriate; 3 were on fluphea~azine (Prolixin) in doses up to 25 mg and 1 was on thiothixene (Navane) up to 30 rag. Six subjects were on Iorazepam (Ativan) PRN throughout the study. Ten subjects were on anticholinergic medication.
All serum samples were frozen at - 7 0 ° until GH determination by RIA and immunoblotic assay. The RIA was performed as described by Sigel et al (1980), using reagents obtained from the National Hormone and Pituitary Program of the NIDDK. The intra and interassay coefficients of variation were less than 10%. The immunoblotic assay (Figure I) was pertbrmed as described by Sinha et al (1984, 1986) with slight modifications (details available upon request).
3
7-
Design
Assays
2
-22K GH -20K GH -17K GH
(,) Immunostalned with antl-hGH Figure I, Autoradiogram of a Western blot analysis of human sera showing the four molecular weight variants of GH measured in this study. Two hundred fifty Ixl of serum samples were used for the analysis. Lane 1 contains recombitant GH 44-191, 10 ng; lane 2, pituitary derived 20K GH, 1Ong; lane 3, 22K GH standard from the NIH, 10 ng; lane 4, serum from a normal individual; and lane 5, serum from a schizophrenic patient. The large band between 22K and 27K GH-immunoreactive bands is the light chain of immunoglobulin molecule, and those in the upper regions of the gel are heavy chain and other unidentified plasma
proteins,
One outlier in the control group had a 17K of 647 and was deleted from the analysis. In order to assess if the GH variants were higher in the schizophrenic group because of elevated RIA GH, a subset of 12 schizophrenic patients with GH levels less than 5 ng/ml were also compared to the normal controls. Both 27K and 22K GH were found to be significantly higher in this subset of 12 patients. All of the above results are in Table ! and the p values shown are after the Bonferroni correction. The raw data for the significant results are seen in Figures 2,3, and 4. No differences were found between the pre and post treatment values for RIA GH and the GH variants in 15 schizophrenic patients as shown in Table 2. Two of the 17 patients did not have a week l blood draw.
Growth Hormone in Schizophrenia
BIOLPSYCHIATRY
489
1993:34:487-491
Table I. Growth Hormone Means and Standard Deviations in 14 Normal Controls, 17 Drug Free Schizophrenic Patients (SCZ), and a Subset of 12 Patients, All with Growth Hormones Less than 5 ng/ml Subjects
RIA GH (ng/ml)
27K
Normal controls n = 14 SCZ drug free n = 17 SCZ drug free n = 12
.77 (.46) 4. I (6) .91 (.57)
q6 (128) 288 (212) ° 284 (220) b
Growth hormene variants (CPM/ml) 22K 20K 52 (104) 292 (300)O 204 (296) b
17K
20 (44) 60 (72) b 52 (72)
108 (120) 256 (168) b 220 (156)
aDiffered from controls p < .01 by Mann-WhitneyU-test bDiffered from controls p < .05 by Mann-WhitneyU-test
760
@
730 670
O
580 1 55(1
O O
1112
@
"/30
0
650
620
0
490
0 o
,:160 @
38(1 I 350
@
410 380
270
O
240
00
:
18o 150
•
60
0
•
O @
90
3O
°
210
O
¢h 180
0 @
150
12o t't,q
240
000 O
'~ 210
®
270
_
O
O O • @ @
00000
--
SCHIZOPHRENIA GROUP
CONTgOl, GROUP
Figure 2. Scattergramof 27K GH variant for all subjects. Open circles refer to patients with GH values greater than 5 ng/ml.
Discussion The schizophrenic subjects had significantly higher levels of the GH variants 27K and 22K, even in the subset of 12 patients whose mean RIA GH levels were approximately equal to the controls. We had previously measured 22K GH and 20K GH in 8 schizophrenic patients and 4 normal controls and found no differences (Peabody 1990). In that study, the sensitivity of the assay used was lower and the sample sizes were small, and thus we could have missed the differences. The clinical significance of our current findings is unknown since little is known about GH variants in schizophrenia. There is no data on GH
O o
L~ 120 ~e e~ 90
3O 0
@
-~ =. °"
@
@ O• O0000000
SCHIZOPHRENI& (|ROUP
(2ONTgOL (]gOUP
Figure 3. Scattergramof 22K GH variant for all subjects. Open circles refer to patients with GH values greater than 5 ng/ml.
variants in other psychiatric disorders, nor is there any known association between GH and psychosis. Differences in GH variants may offer an explanation for the inconsistent results of RIA GH measurements in schizophrenia and thus provide an alternative parameter in assessing GH response. In general, information is unknown or extremely limited regarding GH variants' biological activity, immunological cross-reactivity with one another or the degree of post translational regulation from a pro GH molecule. The 22K GH molecule consists of 191 amino acids and is the predominant form of GH in the pituitary to which all other variants are compared regarding chemical
490
BIOLPSYCHIATRY
M. Dhyanne Warner et al
1993;34"4~7-49!
270 240
•
210
,~
180
O
150
•
. 120 ¢~
9O
o
O
°
•
•
****.****
SCHIZOPHRENIA GROUP
CONTROl, GROUP
Figure 4. Scattergramof 20K GH variant for all subjects. Open circles refer to patients with GH values greater than 5 ng/ml.
structure and biological effect. The 20K variant, with growth "promoting activity similar to 22K and less diabetogenic activity, differs from 22K GH molecule by a deletion of 15 amino acids (Lewis et al 1987), while the i7K GH variant which has no growth promoting activity but more potent glucose intolerance activity in yellow obese mice, may be a fragment of 22K GH consisting of residues 4419 ! (Lewis et al 1991). The identity and biological activity nf the GH-immunoreactive 27K protein is not established, but it may be an alternatively spliced GH molecule having 42 additional residues. The imltmnological cross-reactivity of the 22K variant by RIA is approximately 30% to the 20K variant and less than I% to the 17K variant. Both 20K and I'/K variants are detected in 22K GH antiserum of Western blot, however the degree of their cross-reactivity remains unknown, With few exceptions, investigations with RIA GII in schizophrenia have not been particularly helpful. Baseline serum GH has generally been reported to be within normal limits (Meltzer 1984). GH secretion has been investigated
after apomorphine, L-dopa, insulin-induced hypoglycemia, d-amphetamine, thyrotropin releasing hormone, and leutinizing releasing hormone, and during sleep (Meltzer 1984). These results are either inconsistent or too few studies have been done to make any definitive statements. Recently, we reported a lower GH response to GHRH in patients with schizophrenia (Peabody et al 1990) but 2 other investigators could not confirm this finding (Mayerhoff et al 1990; Nerozzi et al 1990). This pilot data suggests 2 preliminary findings. First, GH variants, 27K and 22K, may be higher in drug free schizophrenic patients. Second, we failed to find any signicant effect on GH variants by neuroleptic treatment. However several confounding variables must be noted. One factor in this study is the difference in sampling times for schizophrenic patients (blood drawn 7:30 a.m.) and normal controls (blood drawn throughout the day), since it is known that RIA GH peaks during sleep. However, in our own previous data we were unable to demonstrate differences in RIA GH or GH variants in morning and afternoon samples in normal controls (Warner et al, in press). It remains possible that it may be different for patients with schizophrenia but no data exist to answer this question. The fact that neuroleptic treatment did not have a significant effect on GH variants may be due to the short drug treatment phase ot ! week which was dictated by hospital time constraints. It is possible that baseline GH measures in these schizophrenics were influenced by a rebound effect from neuroleptic withdrawal; however, two.thirds of the patients were neuroleptic free for an average of 1.6 years and the remaining patients were drug naive. No ongoing measures of anxiety or psychopathology were done and would be useful in future studies to assess stress levels since this is known to effect pituitary hormone secretion, These preliminary data need to be replicated before any definitive conclusions can be made,
The authors wish to gratefully acknowledge Battle Roberts for her excellent technical assistance with the manuscript and Bradley P, Jacobsen for his expertise in performing the Western blot analyses.
Table 2. Growth Hormone Means and Standard Deviations in 15 Drug l~ree ~chizophrenia Subjects (SCZ) Before and I Week After Neuroleptic Medications" Subjecls
RIA GH (nglml)
27K
SCZ baseline SCZ on neurolcptics
3.7 (5,9) 2.3 t4.4)
296 (224) 208 (156)
'q'here were no significanldifferences between groups by paired t-tests.
Growth hormone variants (CPM/ml) 22K 20K 296 (308) 216 (272)
60 (76) 84 (92)
17K 256 (176) 192 (172)
Growth Hormone in Schizophrenia
BIOLPSYCHIATRY 1993;34:487-491
491
References Baumann G (1991): Growth hormone heterogeneity: Genes, isohormones variants, and binding proteins. Endocr Rev 12:424449. Lewis UJ, Lewis LJ, Salelm MAM, Staten NR, Galosy SS, Krivi GG (199 !): A reeombinant-DNA-derived modification of human growth hormone (hGH 44-191) with enhanced diabetogenic activity. Molecular and Cellular Endocrinology 78:4554. Lewis UJ, Markoff E, Culler FL, Hayek A, VanderLaan WP (1987): Biological properties of the 20K-dalton variant of human growth hormone: A review. Endocr Japan 34:73-85. Mayerhoff DI, Lieberman JA, Lemus CA, Pollack S, Schneider BS (1990): Growth hormone response to growth hormonereleasing hormone in schizophrenic patients. Am J Psychiatry 147:1072-1074. Meltzer HY 0984): Neuroendocrine abnormalities in schizophrenia: Prolaetin, growth hormone, and gonadotrophins. In B:own GM, Koslow SH, Reichlin S (eds), Neuroendocrinoiogy in Psychiatric Disorder. New York: Raven Press, pp 1-28. Nerozzi D, Magnani A, Sforza V, Scaramucci E, Cerilli M, Moretti C, Antonozzi !, Frajese G (1990): Prolactin and growth
hormone responses to growth hormone-releasing hormone in acute schizophrvnia. Neuropsychobiology 23:15-17. Peabody CA, Warner MD, Markoff E, Hoffman AR, Wilson DM, Csemansky JG (1990): Growth hormone response to growth hormone releasing hormone in depression and schizophxenia. Psychiatry Res 33:269-276. Sigel MB, VanderLaan WP, VanderLaan EF, Lewis UJ (!980): Measurement of multiple forms of human growth hormone: Cross-reactivities in conventional and two-chain radioimmunassays. Endocrinology 106.92-97. Sinha YN, Gilligan TA, Lee DW (i984): Detection of a high molecular weight variant of prolactin in human plasma by a combination of electrophoretic and immunologic techniques. J Clin Endocrinol Metab 58:752-754, Sinha YN, Gilligan TA, Lee DW, Baxi SC, VanderLaan WP 0986): Demonstration of 20K GH in human plasma by gel electrophoretic-immunostaining-autoradiographic assay (GEISAA). Herin Metab Res 18:402--406. Warner MD, Sinha YN, Peabody CA. Growth hormone and prolactin variants in normal subjects: relative proportions in morning and afternoon samples. Herin Metab Res (in press).
4.87
Potential Abnormalities in Molecular Forms of Growth Hormone in Schizophrenia: A Pilot Study M. Dhyanne Warner, Yagya N. Sinha, and Cecilia A. Peabody
Growth hormone has been investigated in numerous studies involving patients with schizophrenia but has been measured only by radioimmunoassay (RIA). There have been no consistent abnormalities differentiating patients with schizophrenia from normal controls. In the current study, growth hormone (GH ) variants were measured by Western blotting techniques, which resulted in the quantitation of 4 GH size variants: 27K (27,000 Daltons), 22K, 2OK, and i 7K. in the entire sample of ! 7 schizophrenic subjects, all GH variants were significantly higher than in the 14 normal controls. While there were no significant differcnces between the 2 groups in RIA GH values, the RIA values were generally higher in the schizophrenic group. In a subset of 12 schizophrenic patients whose RIA values were approximately equal to the controls, both the 271( and 22K GH variants remained significantly higher in the patient group. In the schizophrenic group, none of the GH variants or RIA GH changed significantly after I week of treatment with neuroleptic medication. These preliminary results suggest that certain GH forms may be elevated in schizophrenia, but further studies are needed. Key Words: Schizophrenia, growth hormone Introduction Numerous studies have investigated growth hormone (GH) in schizophrenia, in part because a dopamine dysregulation theory remains a major pathophysiologic component in schizophrenia, and dopamine stimulates GH. To date there have been no consistent abnormalities to differentiate patients with schizophrenia from normal controls in either basal or stimulated conditions. GH levels have been measured by radioimmunoassay (RIA) only, with the exception of our own previous work (Peabody et al 1990). There have been several molecular forms (variants) of immunoreactive GH identified, with marked differences in molecular weight and biological and immunological activities
From the D e ~ e n t of Psychiatry, StanfordMedical School, Stanford, CA (MDW); The Whittier Institute, La Jolla, CA (YNS); and Department of Fsychiatry, VA Medical Center, Menlo Park, CA (CAP). Address correspondence to Cecilia Peabody MD, V.A, Medical Center, 324-EI03, 795 Willow Road, Menlo Park, CA 94025. Received June 22, 1992; revised June 24, 1993. © 1993 Society of Biological Psychiatry
(Baumatln 1991). Hence, some variants of GH may be abnormal but not detected by standard RIA techniques. Further definition of GH variants may provide some clues to understanding the inconsistent RIA GH results in schizophrenia. In the present study GH was measured by RIA as well as by Western blotting techniques, which resulted in 4 GH size variants identified by molecular weight: 27K, 22K, 20K and 17K. These were measured in normal controls and drug free schizophrenic patients before and 1 week after treatment with neuroleptic medication. Subjects and Methods Subjects Seventeen inpatient men with a DSM-III-R diagnosis of chronic schizophrenic participated in this study. All were at least 2 weeks free from oral and 3 months free from decanoate neuroleptics. The mean age of the schizophrenic 0006-3223/93/$06.00
488
BIOLPSYCHIATRY
M. Dhyanne Warner et al
1993;34:487-491
group was 34 years (range ! 8-47). All subjects were nonobese (< 130% ideal weight), in excellent health, and followed unrestricted dietary regimens. Five patients were drug naive, and 12 patients had been free of neuroleptics for an average of 1.6 years. The average age of onset was 25 years, and the average duration of illness was 8 years. Fourteen normal men served as controls, were in excellent health by history, were on no medication, and had no history of psychiatric illness. The mean age was 32 years (range 23-52). All subjects gave written informed consent after the protocol had been fully explained.
1
Statistics The Mann Whitney U-test was used to determine differences between the controls and schizophrenics since the groups were not normally distributed. A paired t-test was used to determine the differences between the pre- and posttreatment groups since the differences here were. essentially normally distributed. Because of multiple t-tests (5) in this analysis the conservative Bonferroni correction was used and all p values were multiplied by 5.
Results In the entire sample of 17 schizophrenic subjects, all GH variants were significantly higher than in the 14 normal controls but no differences were found for the RIA GH.
4
-27K GH
All inpatient serum samples were drawn at approximately 7:30 a.m. The blood for the normal controls was drawn throughout the day. The patient's blood was again drawn I week after starting medication. Neuroleptics were started after the baseline serum levels were drawn and doses were determined clinically. Thirteen subjects were on haloperidol (Haldol) in doses up to 30 mg per day. In 4 patients haloperidol was not deemed clinically appropriate; 3 were on fluphea~azine (Prolixin) in doses up to 25 mg and 1 was on thiothixene (Navane) up to 30 rag. Six subjects were on Iorazepam (Ativan) PRN throughout the study. Ten subjects were on anticholinergic medication.
All serum samples were frozen at - 7 0 ° until GH determination by RIA and immunoblotic assay. The RIA was performed as described by Sigel et al (1980), using reagents obtained from the National Hormone and Pituitary Program of the NIDDK. The intra and interassay coefficients of variation were less than 10%. The immunoblotic assay (Figure I) was pertbrmed as described by Sinha et al (1984, 1986) with slight modifications (details available upon request).
3
7-
Design
Assays
2
-22K GH -20K GH -17K GH
(,) Immunostalned with antl-hGH Figure I, Autoradiogram of a Western blot analysis of human sera showing the four molecular weight variants of GH measured in this study. Two hundred fifty Ixl of serum samples were used for the analysis. Lane 1 contains recombitant GH 44-191, 10 ng; lane 2, pituitary derived 20K GH, 1Ong; lane 3, 22K GH standard from the NIH, 10 ng; lane 4, serum from a normal individual; and lane 5, serum from a schizophrenic patient. The large band between 22K and 27K GH-immunoreactive bands is the light chain of immunoglobulin molecule, and those in the upper regions of the gel are heavy chain and other unidentified plasma
proteins,
One outlier in the control group had a 17K of 647 and was deleted from the analysis. In order to assess if the GH variants were higher in the schizophrenic group because of elevated RIA GH, a subset of 12 schizophrenic patients with GH levels less than 5 ng/ml were also compared to the normal controls. Both 27K and 22K GH were found to be significantly higher in this subset of 12 patients. All of the above results are in Table ! and the p values shown are after the Bonferroni correction. The raw data for the significant results are seen in Figures 2,3, and 4. No differences were found between the pre and post treatment values for RIA GH and the GH variants in 15 schizophrenic patients as shown in Table 2. Two of the 17 patients did not have a week l blood draw.
Growth Hormone in Schizophrenia
BIOLPSYCHIATRY
489
1993:34:487-491
Table I. Growth Hormone Means and Standard Deviations in 14 Normal Controls, 17 Drug Free Schizophrenic Patients (SCZ), and a Subset of 12 Patients, All with Growth Hormones Less than 5 ng/ml Subjects
RIA GH (ng/ml)
27K
Normal controls n = 14 SCZ drug free n = 17 SCZ drug free n = 12
.77 (.46) 4. I (6) .91 (.57)
q6 (128) 288 (212) ° 284 (220) b
Growth hormene variants (CPM/ml) 22K 20K 52 (104) 292 (300)O 204 (296) b
17K
20 (44) 60 (72) b 52 (72)
108 (120) 256 (168) b 220 (156)
aDiffered from controls p < .01 by Mann-WhitneyU-test bDiffered from controls p < .05 by Mann-WhitneyU-test
760
@
730 670
O
580 1 55(1
O O
1112
@
"/30
0
650
620
0
490
0 o
,:160 @
38(1 I 350
@
410 380
270
O
240
00
:
18o 150
•
60
0
•
O @
90
3O
°
210
O
¢h 180
0 @
150
12o t't,q
240
000 O
'~ 210
®
270
_
O
O O • @ @
00000
--
SCHIZOPHRENIA GROUP
CONTgOl, GROUP
Figure 2. Scattergramof 27K GH variant for all subjects. Open circles refer to patients with GH values greater than 5 ng/ml.
Discussion The schizophrenic subjects had significantly higher levels of the GH variants 27K and 22K, even in the subset of 12 patients whose mean RIA GH levels were approximately equal to the controls. We had previously measured 22K GH and 20K GH in 8 schizophrenic patients and 4 normal controls and found no differences (Peabody 1990). In that study, the sensitivity of the assay used was lower and the sample sizes were small, and thus we could have missed the differences. The clinical significance of our current findings is unknown since little is known about GH variants in schizophrenia. There is no data on GH
O o
L~ 120 ~e e~ 90
3O 0
@
-~ =. °"
@
@ O• O0000000
SCHIZOPHRENI& (|ROUP
(2ONTgOL (]gOUP
Figure 3. Scattergramof 22K GH variant for all subjects. Open circles refer to patients with GH values greater than 5 ng/ml.
variants in other psychiatric disorders, nor is there any known association between GH and psychosis. Differences in GH variants may offer an explanation for the inconsistent results of RIA GH measurements in schizophrenia and thus provide an alternative parameter in assessing GH response. In general, information is unknown or extremely limited regarding GH variants' biological activity, immunological cross-reactivity with one another or the degree of post translational regulation from a pro GH molecule. The 22K GH molecule consists of 191 amino acids and is the predominant form of GH in the pituitary to which all other variants are compared regarding chemical
490
BIOLPSYCHIATRY
M. Dhyanne Warner et al
1993;34"4~7-49!
270 240
•
210
,~
180
O
150
•
. 120 ¢~
9O
o
O
°
•
•
****.****
SCHIZOPHRENIA GROUP
CONTROl, GROUP
Figure 4. Scattergramof 20K GH variant for all subjects. Open circles refer to patients with GH values greater than 5 ng/ml.
structure and biological effect. The 20K variant, with growth "promoting activity similar to 22K and less diabetogenic activity, differs from 22K GH molecule by a deletion of 15 amino acids (Lewis et al 1987), while the i7K GH variant which has no growth promoting activity but more potent glucose intolerance activity in yellow obese mice, may be a fragment of 22K GH consisting of residues 4419 ! (Lewis et al 1991). The identity and biological activity nf the GH-immunoreactive 27K protein is not established, but it may be an alternatively spliced GH molecule having 42 additional residues. The imltmnological cross-reactivity of the 22K variant by RIA is approximately 30% to the 20K variant and less than I% to the 17K variant. Both 20K and I'/K variants are detected in 22K GH antiserum of Western blot, however the degree of their cross-reactivity remains unknown, With few exceptions, investigations with RIA GII in schizophrenia have not been particularly helpful. Baseline serum GH has generally been reported to be within normal limits (Meltzer 1984). GH secretion has been investigated
after apomorphine, L-dopa, insulin-induced hypoglycemia, d-amphetamine, thyrotropin releasing hormone, and leutinizing releasing hormone, and during sleep (Meltzer 1984). These results are either inconsistent or too few studies have been done to make any definitive statements. Recently, we reported a lower GH response to GHRH in patients with schizophrenia (Peabody et al 1990) but 2 other investigators could not confirm this finding (Mayerhoff et al 1990; Nerozzi et al 1990). This pilot data suggests 2 preliminary findings. First, GH variants, 27K and 22K, may be higher in drug free schizophrenic patients. Second, we failed to find any signicant effect on GH variants by neuroleptic treatment. However several confounding variables must be noted. One factor in this study is the difference in sampling times for schizophrenic patients (blood drawn 7:30 a.m.) and normal controls (blood drawn throughout the day), since it is known that RIA GH peaks during sleep. However, in our own previous data we were unable to demonstrate differences in RIA GH or GH variants in morning and afternoon samples in normal controls (Warner et al, in press). It remains possible that it may be different for patients with schizophrenia but no data exist to answer this question. The fact that neuroleptic treatment did not have a significant effect on GH variants may be due to the short drug treatment phase ot ! week which was dictated by hospital time constraints. It is possible that baseline GH measures in these schizophrenics were influenced by a rebound effect from neuroleptic withdrawal; however, two.thirds of the patients were neuroleptic free for an average of 1.6 years and the remaining patients were drug naive. No ongoing measures of anxiety or psychopathology were done and would be useful in future studies to assess stress levels since this is known to effect pituitary hormone secretion, These preliminary data need to be replicated before any definitive conclusions can be made,
The authors wish to gratefully acknowledge Battle Roberts for her excellent technical assistance with the manuscript and Bradley P, Jacobsen for his expertise in performing the Western blot analyses.
Table 2. Growth Hormone Means and Standard Deviations in 15 Drug l~ree ~chizophrenia Subjects (SCZ) Before and I Week After Neuroleptic Medications" Subjecls
RIA GH (nglml)
27K
SCZ baseline SCZ on neurolcptics
3.7 (5,9) 2.3 t4.4)
296 (224) 208 (156)
'q'here were no significanldifferences between groups by paired t-tests.
Growth hormone variants (CPM/ml) 22K 20K 296 (308) 216 (272)
60 (76) 84 (92)
17K 256 (176) 192 (172)
Growth Hormone in Schizophrenia
BIOLPSYCHIATRY 1993;34:487-491
491
References Baumann G (1991): Growth hormone heterogeneity: Genes, isohormones variants, and binding proteins. Endocr Rev 12:424449. Lewis UJ, Lewis LJ, Salelm MAM, Staten NR, Galosy SS, Krivi GG (199 !): A reeombinant-DNA-derived modification of human growth hormone (hGH 44-191) with enhanced diabetogenic activity. Molecular and Cellular Endocrinology 78:4554. Lewis UJ, Markoff E, Culler FL, Hayek A, VanderLaan WP (1987): Biological properties of the 20K-dalton variant of human growth hormone: A review. Endocr Japan 34:73-85. Mayerhoff DI, Lieberman JA, Lemus CA, Pollack S, Schneider BS (1990): Growth hormone response to growth hormonereleasing hormone in schizophrenic patients. Am J Psychiatry 147:1072-1074. Meltzer HY 0984): Neuroendocrine abnormalities in schizophrenia: Prolaetin, growth hormone, and gonadotrophins. In B:own GM, Koslow SH, Reichlin S (eds), Neuroendocrinoiogy in Psychiatric Disorder. New York: Raven Press, pp 1-28. Nerozzi D, Magnani A, Sforza V, Scaramucci E, Cerilli M, Moretti C, Antonozzi !, Frajese G (1990): Prolactin and growth
hormone responses to growth hormone-releasing hormone in acute schizophrvnia. Neuropsychobiology 23:15-17. Peabody CA, Warner MD, Markoff E, Hoffman AR, Wilson DM, Csemansky JG (1990): Growth hormone response to growth hormone releasing hormone in depression and schizophxenia. Psychiatry Res 33:269-276. Sigel MB, VanderLaan WP, VanderLaan EF, Lewis UJ (!980): Measurement of multiple forms of human growth hormone: Cross-reactivities in conventional and two-chain radioimmunassays. Endocrinology 106.92-97. Sinha YN, Gilligan TA, Lee DW (i984): Detection of a high molecular weight variant of prolactin in human plasma by a combination of electrophoretic and immunologic techniques. J Clin Endocrinol Metab 58:752-754, Sinha YN, Gilligan TA, Lee DW, Baxi SC, VanderLaan WP 0986): Demonstration of 20K GH in human plasma by gel electrophoretic-immunostaining-autoradiographic assay (GEISAA). Herin Metab Res 18:402--406. Warner MD, Sinha YN, Peabody CA. Growth hormone and prolactin variants in normal subjects: relative proportions in morning and afternoon samples. Herin Metab Res (in press).