BIOL PSYCHIATRY
1035
1992;32:1035-1041
Elevated Levels of Acute Phase Plasma Proteins in Major Depression Peter R. Joyce, Charles R. Hawes, Roger T. Mulder, J. Douglas Sellman, Deborah A. Wilson, and D. Ross Boswell
Levels of acute phase and other plasma proteins were measured in 21 men with major depression, 28 men with alcohol dependence, and 12 men who acted as controls. The depressed men had significantly elevated levels of the acute phase proteins, haptogloblin and alpha-l-antichymotrypsin, and of immunoglobulin G. The elevations in haptoglobin and alpha-l-antichymotrypsin were highly correlated with each other, and were correlated with the severity of depression and negatively correlated with the thyroid stimulating hormone response to thyrotropin. The alcoholic men had elevated haptoglobin levels, but significantly decreased levels of immunoglobulin G. These findings provide further evidence for an inflammatory response during depression.
Introduction Nemeroff and colleagues (1990) have recently reported that depressed patients have elevated levels of the plasma protein, alpha- l-acid glycoprotein. The interest in this finding is that alpha-l-acid glycoprotein has been identified as an endogenous modulator of the tritiated imipramine/5HT transporter in human platelets (Abraham et al 1987), and that alterations of this endogenous inhibitor may be of relevance to the amine hypothesis of depression. Furthermore, Nemeroff et al (1990) found that alpha-l-acid glycoprotein was positively correlated with severity of depression and postdexamethasone plasma cortisol concentrations. However, Nemeroffet al measured only alpha-l-acid glycoprotein (AGP), despite the fact that this protein is one of a series of acute phase proteins produced in the liver, which include alpha-l-antichymotrypsin (AAC), alpha-]-antitrypsin (AAT), haptoglobin (HAFT,), ceruloplasmin (CPM), and C-reactive protein (CRP) (positive acute phase proteins). In addition, an acute phase plasma protein response is usually associated with a decrease in the proteins prealbumin (PREALB) and transferrin (TRANS) (negative acute phase proteins) (Laurell 1977; Alper 1974; Kushner 1988). Subsequently, Healy et al (1991) and Kehoe et al (1991) have replicated the finding of elevated alpha-l-acid glycoprotein levels in depressed patients. However, Maes et al (1992) have observed that depressed patients have increases in other acute phase plasma proteins, including hap-
From the University Department of Psychological Medicine (PP,J, RTM, JDS, DAW), Christchurch School of Medicine; the Department of C~inicai .~iochemistry (CRH, DRB), Christchurch Hospital; and the University Department of Pathology (DBB), Chrit;tchurch School of Medicine, Christchurch, New Zealand. Address reprint r,~xlueststo Professor Peter R. Joyce, Department of Psychological Medicine, Christchurch School of Medicine, P.O. Box 4345, Christchurch, N2::° Zealand. Received April 23, 1991; August 13, 1992. © 1992 Society of Biological Psychiatry
0006-3223/92/$05.00
1036
BIOL PSYCHIATRY 1992;32:1035-~041
P.R. Joyce et al
toglobin and alpha-l-antitrypsin, and decreases in transferrin and albumin, but they were unable to replicate findings of elevated alpha-l-acid glycoprotein levels. Maes et al (1992) intertJret these findings, along with other immune changes, as evidence for an inflammatory response in depressive disorders. The purpose of the present study was to measure a range of plasma proteins, including alpha- l-acid glycoprotein, haptoglobin, alpha- l-antitrypsin, and other acute phase plasma proteins in depressed and healthy men. As a further control group we also measured these plasma proteins in a group of alcoholic men, to examine the specificity of any changes to depression. Methods
Subjects The depressed patients and control subjects for this study were recruited as part of a study on the prediction of antidepressant drug response, although the alcoholic men were recruited as part of a study on relapse in alcoholism. All subjects were required to given written informed consent. Subjects were physically healthy and had screening tests for thyroid function, liver function, renal function, and blood counts. Subjects were required to be free of medication for a minimum of 2 weeks, or longer for drugs such as anticonvulsants, and fluoxetine; however, the majority of depressed patients and controls had never used psychotropic drugs, and had been drug free for considerably longer than 2 weeks if they had previously used medications. The alcoholic men had been abstinent from any alcohol for a minimum of 3 weeks. All subjects were interviewed by a psychiatrist using the Structured Clinical Interview for DSM-III-R (SCID). The SCID was expanded for the purpose of the present study to include an elaboration as to why patients met any criteria for major depression (e.g., the psychiatrist had to rate whether a patient had decreased appetite, increased appetite, weight loss or weight gain; or whether a patient had ~houghts of death, suicidal ideation, or suicide attempt), and was rated on each of the DSM-III-R melancholic criteria. The dept'essed and control subjects were also rated by the same psychiatrist on the 17 item Hamilton (1960) depression rating scale. All depressed patients received a principal diagnosis of major depressive disorder; all alcoholics received a principal diagnosis of alcohol dependence; and none of the controls had a current psychiatric diagnosis, although one had a past history of major depression, and one had a past history of drug abuse.
Neurobiologic Assessment On the first morning of the neurobiologic assessment the subjects reported to the Clinical Research Unit at 8:30 AM after an overnight fast. Subjects were weighted and measured, and then rested supine on a bed. An indwelling intravenous catheter was inserted into an antecubital vein, which was kept patent with heparinized saline. Blood was drawn for the measurement of glucose, electrolytes, liver function, blood count, and thyroid function tests. Further blood was drawn at 9 AM, which was used for serum proteins, cortisol, prolactin, and growth hormone, prior to receiving oral fenfluramine for a neuroendocrine challenge test. On a second afternoon subjects again reported to the Clinical Research Unit for further neuroendocrine assessment. Blood samples were drawn for plasma cortisol at 15-20 mir~
Acute Phase Plasma Proteins in Depression
BIOL PSYCHIATRY 1992;32:1035-1041
1037
intervals for the 3 hrs from 1 PM-5 PM, with 400 ttg of thyrotropin releasing hormone (TRH) being infused at 3 PM. Thyroid stimulating hormone was measured prior to and following the TRH. Blood samples were centrifuged and frozen for later assay.
Hormone and Protein Assays Specific protein concentrations were measured by electroimmunoassay (Laurell rockets) (Laurell 1972) using antisera purchased from Dakopatts a/s (Copenhagen, Denmark). Results for alpha-l-antichymotrypsin (AAC), alpha-l-antitrypsin (AAT), haptoglobin (HALT), alpha-l-acid glycoprotein (AGP), and ceruloplasmin (CPM) were referred to a control pool from 200 normal blood donors, which was defined as 100%. Results for prealbumin, transferrin, immunoglobulin G (IgG), immunoglobulin A (IgA), and immunoglobulin M (IgM) are expressed as g/L, and results for C-reactive protein are expressed as mg/L. Cortisol was measured by an enzyme-linked immunosorbent assay (Lewis and Elder 1985), which has an interassay coefficient of variation of less than 11%. Serum TSH was measured using radioimmunoassay, which is standardized against the World Health Organization standard (801558) and has an interassay coefficient of variation of 15%. The mean afternoon cortisol (CORTPM) was calculated by averaging the eight afternoon cortisol levels and is expressed as nmol/L. Delta TSH was calculated by subtracting the baseline TSH from the peak TSH level at 15, 30, or 45 rain, and are presented as mlU/L.
Statistical Analysis All data were entered into the relational data base Paradox (version 3), and were transferred to SYSTAT for statistical analysis using analysis of variance, post hoc Tukey tests, multiple regression, and Pearson correlations. Results Table 1 shows the plasma levels of all the plasma proteins measured in the 21 men with a current major depressive disorder, 28 men with alcohol dependence, and the 12 cow trol men (who suffered from no current psychiatric disorder). An analysis of variance (ANOVA) across the three groups of men revealed that the groups differed in terms of transferrin (TRANS) levels (F = 4.44, p = 0.016), alpha-l-antichymotrypsin (AAC) ( F = 6 . 7 3 , p - 0.002), haptoglobin (HAPT) (F = 3.88, p = 0.026), ceruloplasmin (CRM) ( F - 5.79, p = 0.005), and immunoglobulin G (F = 26.2, p = 0.000). Post hoc Tukey tests showed that the depressed men had significantly elevated levels of alphal-antichymotrypsin (p - 0.004), haptoglobin (p - 0.008), and IgG (p - 0.023) compared with the controls. The alcohol dependent men had elevated levels of haptoglobin (p = 0.014) and significantly lower levels of IgG (p = 0.002). Table 2 shows the correlation matrix between the plasma proteins, Hamilton Depression Rating Score (HDRS), mean afternoon cortisol, and the TSH response to TRH (DTSH) in the depressed men. From Table 2 it can be seefl that the levels of acute positive phase proteins were all positively correlated with the Pearson correlations between AAC, AAT, HALT, and AGP ranging from 0.52 to 0.80 (all p's < 0.01). The highest correlation was between AAC and HALT. These two acute phase plasma proteins were not significantly correlated with the negative acute phase protein,,; c~r with IgG, but both were also
1038
P.R. Joyce et al
BIOL PSYCHIATRY 1992;32:1035-1041
Table 1. The Mean Levels (_+ SD) of Plasma Proteins in Depressed, Conu-ol .andAlcoholic Men n
21 Depressives 34.2 23.1 280 14.9 20.6
Age Body mass index Mean PM cortisol
Delta TSH Hamilton score Plasma Proteins
12 Controls
(10.1) (2.9) (79) (8.5) (3.1)
31.5 24.2 292 22.4 1.5
Depressives
PREALB TRANS AAC AAT HAFT AGP CPM CRP IGG IGA IGM
0.26 2.70 101.7 84.0 128. I 90.0 82.2 2.9 12.2 2.41 1.18
(4.1) (2.9) (112) (14.0) (1.3)
28 Alcoholics 33.7 25.4 261 20.3
Controls
(0.07) (0.34) (21.0) (18.6) (55.7) (22.9) (16.2) (3.7) (I.7) (0,94) (0,47)
0.27 2,59 79.4 78.1 78.6 81.1 73.8 4.3 10,7 !.96 1,27
(0.07) (0.23) (17.1) (7.0) (29.8) (13.9) (20.2) (3.5) (I.9) (0.68) (0.64)
(8.9) (3.3) (88) (17.4) ANOVA p
Alcoholics
0.25 2.43 92.4 82.5 118.0 87.8 70.9 3.0 8.9 1.86 !.27
(0,04) (0,34) (23.1) (14.2) (49.4) (22.5) (13.3) (3.9) (I.4) (0.76) (0,50)
NS 0.016 0.002 NS 0.026 NS 0.005 NS 0.000 NS NS
significantly correlated with the Hamilton Depression Rating Score (p < 0.05), and were negatively correlated with the TSH response to TRH infusion. Neither was significantly correlated with the mean afternoon cortisol level. In the depressed men, a ~ultiple regression on either alpha- l-chymotrypsin or haptoglobin levels revealed that the Hamilton score was no longer a significant correlate when the TSH response to TRH was included in the regression.
Discussion The major finding from thi~ study is that depressed patients do have elevated levels of a number of acute phase plasr~mproteins comp~red with controls. In this study the elevations of alpha-l-chymotrypsin and haptoglobin ~vere significa;~t, and there was a trend for
Table 2. Pearson Correlations between Plasma Proteins, Endocrine Variables and Depression Severity in Depressed Men AAT AAC AAT HAPT AGP CPM CRP PREALB TRANS IGG HDRS CORTPM
0,60
HAPT 0.80 0,7_,.22
AGP
CPM,
0,7_..33 0,58 0,5..~2 0,24 0,63 0,66 0,6"/
CRP
PREALB
TRANS
IGG
HDRS
CORTPM
DTSH
0,59 0,58 0,51 0,34 0,08
- 0,04 - 0,27 0,00 0.11 0,17 -0,36
0,13 0,06 0.27 0.21 0,27 -0.14 0,47
- 0.18 - 0.19 -0.38 -0.24 -0,23 -0.06 - 0.24 0.18
0_4,33 0.17 0.54, 0.14 0.18 0,08 0.18 0.21 0.01
U. 12 - ft. 17 024 0.16 0.50 -0.18 0.41 0.01 - 0.19 0.43
- 9.55 - 0.40 -0.43 -0.30 -0.07 0.08 - 0.09 -6.34 0.02 - 0.05 0.16
For n -- 21; if • > 0.42, p < 0.05; • > 0.53, p < 0.01; • > 0.65, p < 0.001 Underline • > ~.42.
Acute Phase Plasma Proteins in Depression
BIOL PSYCHIATRY 1992;32:1035-1041
1039
elevations in other acute phase proteins including alpha-l-acid glycoprotein. These findings are thus most comparable to those of Maes et al (1992) who reported elevated levels of haptoglobin and alpha-l-antitrypsin, but not elevations in alpha-l-acid glycoprotein, whereas Nemeroff et al (1990), Kehoe et al (1991), and Healy et al (1991) who only measured alpha-l-acid glycoprotein reported that this acute phase protein was elevated. If our sample had been larger, we may have replicated the increase in alpha-l-acid glycoprotein, but like Maes et al (1992) our findings suggest that there are more marked changes in other acute phase proteins. It is also of note that these depressed men also had significantly elevated levels of immunoglobulin G. These increases were polyclonal in nature. Two recent reports are of interest with regard to the findings from this study. Firstly, Pitts et al (1990) have reported that depressed patients, and especially depressed men have increases in cerebrospinal fluid (CSF) protein levels. They comment that this increase in CSF protein is not of monoclonal origin but appears to represent an increase in albumin and globulin fractions, although levels of individual proteins were not measured. Secondly, Maes et al (1991) have reported that depressed patients have elevated plasma levels of anticardiolipin and antinuclear antibodies, and of interleukin-2 receptors. These authors postulate that these findings could be consistent with modifications in cellular membranes during depressive episodes, and comment on other disease states such as autoimmune disorders, chronic neurological disorders, and infections that also give rise to co.reparable findings. Certainly, when the immunological abnormalities observed in depre~ion are combined with changes in immunoglobulin G and acute phase plasma proteins, it is hard not to agree with Maes et al (1992) that major depression is often accompanied by evidence of an inflammatory process. Both Nemeroff et al (1990) and Healy et al (1991) found that elevations in alpha-lacid glycoprotein were associated with higher postdexamethasone co~isol levels. In this study the elevations of haptoglobin and alpha-l-chymotrypsin were not related to the mean afternoon cortisol level, but both were associated with a blunted TSH response to TRH. This association has not previously been reported. However, in this group of patients we do not have clear evidence of a general cortisol hypersecretion, as assessed by the mean afternoon cortisol level. It needs to be recalled though that elevated mean afternoon cortisol levels and dexamethasone suppression test nonsuppression are not necessarily identical (ttolsboer et al 1984). The interesting question that arises from this research is the nature of the association between elevations in acute phase proteins and blunting of the TSH response to TRH. On present data it is not possible to establish whether the protein changes could be a result of the hypothalamic-pituitary-thyroid axis abnormality, or vice-versa, or whether both are reflections of some other underlying abnormality in depression. In evaluating the timings from this study a number of methodological issues should be noted. All the control men underwent their neurobiological assessment on an outpatient basis, as did 20 of the 21 depressed men. However, the alcoholic men were all inpatients, and had been inpatients for about 3 weeks at the time of their neurobiological assessment. The levels of the acute phase proteins in the control men are generally about 80%, which is perhaps lower than would be expected from a control group given that the assays are standardized to a pooled control plasma giving levels of !q0%. However, a number of issues are of note in this regard. In the first instance the pooled plasma for the assay standard comprises plasma from both men and women, whereas we only measured levels in men so as to avoid any possible effects of gender. The second factor of note is that the blood samples in this study were drawn through an indwelling intravenous catheter,
1040
BIOL PSYCHIATRY 1992;32:1035-1041
P.R. Joyce et al
while the subject rested supine, and without the need for a tourniquet. This would probably lead to lower levels of plasma proteins than if they had been drawn by venipuncture with the use of a tourniquet. This would account for the lower levels of acute phase proteins in our control subjects compared to the reference population. However, it does not invalidate the differences found between our controls and depressed or alcoholic patients, as all the subjects in this study had blood drawn in a standard manner, at the same time of the day, by the same nurse, and blood samples were subsequently handled in a comparable fashion. A third methodological issue is that we have performed multiple analysis of variances and Pearson correlations on the data collected, and have not made any explicit adjustment for having performed multiple statistical tests. However, the significance of many of the findings is well beyond a 0.05 level of significance, and within the general pattern of results there is a consistency to the fir,dings that provides further support for their significance. This study could thus be taken to in part support the initial findings of Nemeroff et al (1990), which prompted this research. However, the elevated levels of alpha-l-acid glycoprotein, are probably only one aspect of a more general metabolic disturbance in plasma proteins in depressed patients. ~f alpha- l-acid glycoprotein is part of this general elevation of acute phase plasma proteihs, and is an inhibitor of plateiet 5HT/imipramine binding site,~ then this could possibly be of relevance to serotonergic mechanisms in depression. However, our findings also raise a whole series of further questions as to why such a generalized increase in a variety of acute phase plasma proteins and immunoglobulin G occurs in major depression, and what are the mechanisms that produce this increase? In considering the possible mechanisms for this generalized elevation of plasma proteins it should be recalled that the depressed patients in this study had been depressed for months or years prior to participating in this study, which also raises the question as to why such an acute phase response can presumably persist over a prolonged period of time, We wish to thank Mrs, lsobel Stevens and Robyn Abbott for their assistance in completing this study. This project was funded by grants from the Medical Research Council of New Zealand and the Alcoholic Liquor
Advisory Council,
References Abraham KI, leni JR, Meyerson LR (1987): Purification and properties of a human plasma endogenous modulator for the platelet tricyclic binding/semtonin transport complex. Biochem Biophys Acta 932:8-21. Alper CA (19"/4): Plasr,:a protein measurements as a diagnostic aid. N Engl J Med 291:287-291. Hamilton M (1960): A rating scale for depression. J Neurol Neurosurg Psychiatry 23:56-62. Healy D, Calvin J, Whitehouse AM, et al (19~i): Alpha-i-acid glycoprotein in major depression and eating disorders. J Affective Disord 22:13-20. Holsboer F, Gerken A, Steiger A, Fass V (1984): Mean 1400-1700 plasma cortisol concentration and its relationship to the Img dexamethasone suppression response in depressions and controls. Acta Psychiatr Scand 69:383-390. Kehoe WA, Kwentus JA, Sheffel WB, Harralson AF (1991): Increased alpha-l-acid glycoprotein in depression lowers free fraction of imipramine. Biol Psychiatry 29:489-493. Kushner ! (1988): The acute phase response: An overview. In Methods in Enzymology. Vol 163, pp 373-383.
Acute Phase Plasma Proteins in Depression
BIOL PSYCHIATRY 1992;32:1035-1041
1041
Laurell CB (1972): Electrophoretic and electro-immunochemical analysis of proteins. Scand J Clin Lab Invest 29:(suppl 124). Lewis JG, Elder PA (1985): An enzyme-linked immunosorbent assay (ELISA) for plasma cortisol. J Steroid Biochem 22:673-676. Maes M, Bosmans E, Suy E, Vandervorst C, Dejonckheere C, Raus J (1991): Antiphospholipid, antinuclear, Epstein-Barr and cytomegalovirus antibodies, and soluble interleukin-2 receptors in depressive patients. J Affective Disord 21:133-140. Maes M, Scharpe S, Van Grootei L, et al (1992): Higher alpha-l-antitrypsin, haptoglobin, ceruloplasmin and lower retinol binding protein plasma levels during depression: Further evidence for the existence of an inflammatory response during that illness. J Affective Disord 24:183192. Nemeroff CB, Krishnan KRR, Blazer DO, Knight DL, Benjamin D, Meyerson LR (1990): Elevated plasma concentrations of alpha-l-acid glycoprotein, a putative endogenous inhibitor of the tritiated imipramine binding site, in depressed patients. Arch Gen Psychiatry 47:337-340. Pitts AF, Carroll BT, Gehris TL, Kathol RG, Samuelson SD (1990): Elevated CSF protein in male patients with depression. Bioi Psychiatry 28:629-637.
1035
1992;32:1035-1041
Elevated Levels of Acute Phase Plasma Proteins in Major Depression Peter R. Joyce, Charles R. Hawes, Roger T. Mulder, J. Douglas Sellman, Deborah A. Wilson, and D. Ross Boswell
Levels of acute phase and other plasma proteins were measured in 21 men with major depression, 28 men with alcohol dependence, and 12 men who acted as controls. The depressed men had significantly elevated levels of the acute phase proteins, haptogloblin and alpha-l-antichymotrypsin, and of immunoglobulin G. The elevations in haptoglobin and alpha-l-antichymotrypsin were highly correlated with each other, and were correlated with the severity of depression and negatively correlated with the thyroid stimulating hormone response to thyrotropin. The alcoholic men had elevated haptoglobin levels, but significantly decreased levels of immunoglobulin G. These findings provide further evidence for an inflammatory response during depression.
Introduction Nemeroff and colleagues (1990) have recently reported that depressed patients have elevated levels of the plasma protein, alpha- l-acid glycoprotein. The interest in this finding is that alpha-l-acid glycoprotein has been identified as an endogenous modulator of the tritiated imipramine/5HT transporter in human platelets (Abraham et al 1987), and that alterations of this endogenous inhibitor may be of relevance to the amine hypothesis of depression. Furthermore, Nemeroff et al (1990) found that alpha-l-acid glycoprotein was positively correlated with severity of depression and postdexamethasone plasma cortisol concentrations. However, Nemeroffet al measured only alpha-l-acid glycoprotein (AGP), despite the fact that this protein is one of a series of acute phase proteins produced in the liver, which include alpha-l-antichymotrypsin (AAC), alpha-]-antitrypsin (AAT), haptoglobin (HAFT,), ceruloplasmin (CPM), and C-reactive protein (CRP) (positive acute phase proteins). In addition, an acute phase plasma protein response is usually associated with a decrease in the proteins prealbumin (PREALB) and transferrin (TRANS) (negative acute phase proteins) (Laurell 1977; Alper 1974; Kushner 1988). Subsequently, Healy et al (1991) and Kehoe et al (1991) have replicated the finding of elevated alpha-l-acid glycoprotein levels in depressed patients. However, Maes et al (1992) have observed that depressed patients have increases in other acute phase plasma proteins, including hap-
From the University Department of Psychological Medicine (PP,J, RTM, JDS, DAW), Christchurch School of Medicine; the Department of C~inicai .~iochemistry (CRH, DRB), Christchurch Hospital; and the University Department of Pathology (DBB), Chrit;tchurch School of Medicine, Christchurch, New Zealand. Address reprint r,~xlueststo Professor Peter R. Joyce, Department of Psychological Medicine, Christchurch School of Medicine, P.O. Box 4345, Christchurch, N2::° Zealand. Received April 23, 1991; August 13, 1992. © 1992 Society of Biological Psychiatry
0006-3223/92/$05.00
1036
BIOL PSYCHIATRY 1992;32:1035-~041
P.R. Joyce et al
toglobin and alpha-l-antitrypsin, and decreases in transferrin and albumin, but they were unable to replicate findings of elevated alpha-l-acid glycoprotein levels. Maes et al (1992) intertJret these findings, along with other immune changes, as evidence for an inflammatory response in depressive disorders. The purpose of the present study was to measure a range of plasma proteins, including alpha- l-acid glycoprotein, haptoglobin, alpha- l-antitrypsin, and other acute phase plasma proteins in depressed and healthy men. As a further control group we also measured these plasma proteins in a group of alcoholic men, to examine the specificity of any changes to depression. Methods
Subjects The depressed patients and control subjects for this study were recruited as part of a study on the prediction of antidepressant drug response, although the alcoholic men were recruited as part of a study on relapse in alcoholism. All subjects were required to given written informed consent. Subjects were physically healthy and had screening tests for thyroid function, liver function, renal function, and blood counts. Subjects were required to be free of medication for a minimum of 2 weeks, or longer for drugs such as anticonvulsants, and fluoxetine; however, the majority of depressed patients and controls had never used psychotropic drugs, and had been drug free for considerably longer than 2 weeks if they had previously used medications. The alcoholic men had been abstinent from any alcohol for a minimum of 3 weeks. All subjects were interviewed by a psychiatrist using the Structured Clinical Interview for DSM-III-R (SCID). The SCID was expanded for the purpose of the present study to include an elaboration as to why patients met any criteria for major depression (e.g., the psychiatrist had to rate whether a patient had decreased appetite, increased appetite, weight loss or weight gain; or whether a patient had ~houghts of death, suicidal ideation, or suicide attempt), and was rated on each of the DSM-III-R melancholic criteria. The dept'essed and control subjects were also rated by the same psychiatrist on the 17 item Hamilton (1960) depression rating scale. All depressed patients received a principal diagnosis of major depressive disorder; all alcoholics received a principal diagnosis of alcohol dependence; and none of the controls had a current psychiatric diagnosis, although one had a past history of major depression, and one had a past history of drug abuse.
Neurobiologic Assessment On the first morning of the neurobiologic assessment the subjects reported to the Clinical Research Unit at 8:30 AM after an overnight fast. Subjects were weighted and measured, and then rested supine on a bed. An indwelling intravenous catheter was inserted into an antecubital vein, which was kept patent with heparinized saline. Blood was drawn for the measurement of glucose, electrolytes, liver function, blood count, and thyroid function tests. Further blood was drawn at 9 AM, which was used for serum proteins, cortisol, prolactin, and growth hormone, prior to receiving oral fenfluramine for a neuroendocrine challenge test. On a second afternoon subjects again reported to the Clinical Research Unit for further neuroendocrine assessment. Blood samples were drawn for plasma cortisol at 15-20 mir~
Acute Phase Plasma Proteins in Depression
BIOL PSYCHIATRY 1992;32:1035-1041
1037
intervals for the 3 hrs from 1 PM-5 PM, with 400 ttg of thyrotropin releasing hormone (TRH) being infused at 3 PM. Thyroid stimulating hormone was measured prior to and following the TRH. Blood samples were centrifuged and frozen for later assay.
Hormone and Protein Assays Specific protein concentrations were measured by electroimmunoassay (Laurell rockets) (Laurell 1972) using antisera purchased from Dakopatts a/s (Copenhagen, Denmark). Results for alpha-l-antichymotrypsin (AAC), alpha-l-antitrypsin (AAT), haptoglobin (HALT), alpha-l-acid glycoprotein (AGP), and ceruloplasmin (CPM) were referred to a control pool from 200 normal blood donors, which was defined as 100%. Results for prealbumin, transferrin, immunoglobulin G (IgG), immunoglobulin A (IgA), and immunoglobulin M (IgM) are expressed as g/L, and results for C-reactive protein are expressed as mg/L. Cortisol was measured by an enzyme-linked immunosorbent assay (Lewis and Elder 1985), which has an interassay coefficient of variation of less than 11%. Serum TSH was measured using radioimmunoassay, which is standardized against the World Health Organization standard (801558) and has an interassay coefficient of variation of 15%. The mean afternoon cortisol (CORTPM) was calculated by averaging the eight afternoon cortisol levels and is expressed as nmol/L. Delta TSH was calculated by subtracting the baseline TSH from the peak TSH level at 15, 30, or 45 rain, and are presented as mlU/L.
Statistical Analysis All data were entered into the relational data base Paradox (version 3), and were transferred to SYSTAT for statistical analysis using analysis of variance, post hoc Tukey tests, multiple regression, and Pearson correlations. Results Table 1 shows the plasma levels of all the plasma proteins measured in the 21 men with a current major depressive disorder, 28 men with alcohol dependence, and the 12 cow trol men (who suffered from no current psychiatric disorder). An analysis of variance (ANOVA) across the three groups of men revealed that the groups differed in terms of transferrin (TRANS) levels (F = 4.44, p = 0.016), alpha-l-antichymotrypsin (AAC) ( F = 6 . 7 3 , p - 0.002), haptoglobin (HAPT) (F = 3.88, p = 0.026), ceruloplasmin (CRM) ( F - 5.79, p = 0.005), and immunoglobulin G (F = 26.2, p = 0.000). Post hoc Tukey tests showed that the depressed men had significantly elevated levels of alphal-antichymotrypsin (p - 0.004), haptoglobin (p - 0.008), and IgG (p - 0.023) compared with the controls. The alcohol dependent men had elevated levels of haptoglobin (p = 0.014) and significantly lower levels of IgG (p = 0.002). Table 2 shows the correlation matrix between the plasma proteins, Hamilton Depression Rating Score (HDRS), mean afternoon cortisol, and the TSH response to TRH (DTSH) in the depressed men. From Table 2 it can be seefl that the levels of acute positive phase proteins were all positively correlated with the Pearson correlations between AAC, AAT, HALT, and AGP ranging from 0.52 to 0.80 (all p's < 0.01). The highest correlation was between AAC and HALT. These two acute phase plasma proteins were not significantly correlated with the negative acute phase protein,,; c~r with IgG, but both were also
1038
P.R. Joyce et al
BIOL PSYCHIATRY 1992;32:1035-1041
Table 1. The Mean Levels (_+ SD) of Plasma Proteins in Depressed, Conu-ol .andAlcoholic Men n
21 Depressives 34.2 23.1 280 14.9 20.6
Age Body mass index Mean PM cortisol
Delta TSH Hamilton score Plasma Proteins
12 Controls
(10.1) (2.9) (79) (8.5) (3.1)
31.5 24.2 292 22.4 1.5
Depressives
PREALB TRANS AAC AAT HAFT AGP CPM CRP IGG IGA IGM
0.26 2.70 101.7 84.0 128. I 90.0 82.2 2.9 12.2 2.41 1.18
(4.1) (2.9) (112) (14.0) (1.3)
28 Alcoholics 33.7 25.4 261 20.3
Controls
(0.07) (0.34) (21.0) (18.6) (55.7) (22.9) (16.2) (3.7) (I.7) (0,94) (0,47)
0.27 2,59 79.4 78.1 78.6 81.1 73.8 4.3 10,7 !.96 1,27
(0.07) (0.23) (17.1) (7.0) (29.8) (13.9) (20.2) (3.5) (I.9) (0.68) (0.64)
(8.9) (3.3) (88) (17.4) ANOVA p
Alcoholics
0.25 2.43 92.4 82.5 118.0 87.8 70.9 3.0 8.9 1.86 !.27
(0,04) (0,34) (23.1) (14.2) (49.4) (22.5) (13.3) (3.9) (I.4) (0.76) (0,50)
NS 0.016 0.002 NS 0.026 NS 0.005 NS 0.000 NS NS
significantly correlated with the Hamilton Depression Rating Score (p < 0.05), and were negatively correlated with the TSH response to TRH infusion. Neither was significantly correlated with the mean afternoon cortisol level. In the depressed men, a ~ultiple regression on either alpha- l-chymotrypsin or haptoglobin levels revealed that the Hamilton score was no longer a significant correlate when the TSH response to TRH was included in the regression.
Discussion The major finding from thi~ study is that depressed patients do have elevated levels of a number of acute phase plasr~mproteins comp~red with controls. In this study the elevations of alpha-l-chymotrypsin and haptoglobin ~vere significa;~t, and there was a trend for
Table 2. Pearson Correlations between Plasma Proteins, Endocrine Variables and Depression Severity in Depressed Men AAT AAC AAT HAPT AGP CPM CRP PREALB TRANS IGG HDRS CORTPM
0,60
HAPT 0.80 0,7_,.22
AGP
CPM,
0,7_..33 0,58 0,5..~2 0,24 0,63 0,66 0,6"/
CRP
PREALB
TRANS
IGG
HDRS
CORTPM
DTSH
0,59 0,58 0,51 0,34 0,08
- 0,04 - 0,27 0,00 0.11 0,17 -0,36
0,13 0,06 0.27 0.21 0,27 -0.14 0,47
- 0.18 - 0.19 -0.38 -0.24 -0,23 -0.06 - 0.24 0.18
0_4,33 0.17 0.54, 0.14 0.18 0,08 0.18 0.21 0.01
U. 12 - ft. 17 024 0.16 0.50 -0.18 0.41 0.01 - 0.19 0.43
- 9.55 - 0.40 -0.43 -0.30 -0.07 0.08 - 0.09 -6.34 0.02 - 0.05 0.16
For n -- 21; if • > 0.42, p < 0.05; • > 0.53, p < 0.01; • > 0.65, p < 0.001 Underline • > ~.42.
Acute Phase Plasma Proteins in Depression
BIOL PSYCHIATRY 1992;32:1035-1041
1039
elevations in other acute phase proteins including alpha-l-acid glycoprotein. These findings are thus most comparable to those of Maes et al (1992) who reported elevated levels of haptoglobin and alpha-l-antitrypsin, but not elevations in alpha-l-acid glycoprotein, whereas Nemeroff et al (1990), Kehoe et al (1991), and Healy et al (1991) who only measured alpha-l-acid glycoprotein reported that this acute phase protein was elevated. If our sample had been larger, we may have replicated the increase in alpha-l-acid glycoprotein, but like Maes et al (1992) our findings suggest that there are more marked changes in other acute phase proteins. It is also of note that these depressed men also had significantly elevated levels of immunoglobulin G. These increases were polyclonal in nature. Two recent reports are of interest with regard to the findings from this study. Firstly, Pitts et al (1990) have reported that depressed patients, and especially depressed men have increases in cerebrospinal fluid (CSF) protein levels. They comment that this increase in CSF protein is not of monoclonal origin but appears to represent an increase in albumin and globulin fractions, although levels of individual proteins were not measured. Secondly, Maes et al (1991) have reported that depressed patients have elevated plasma levels of anticardiolipin and antinuclear antibodies, and of interleukin-2 receptors. These authors postulate that these findings could be consistent with modifications in cellular membranes during depressive episodes, and comment on other disease states such as autoimmune disorders, chronic neurological disorders, and infections that also give rise to co.reparable findings. Certainly, when the immunological abnormalities observed in depre~ion are combined with changes in immunoglobulin G and acute phase plasma proteins, it is hard not to agree with Maes et al (1992) that major depression is often accompanied by evidence of an inflammatory process. Both Nemeroff et al (1990) and Healy et al (1991) found that elevations in alpha-lacid glycoprotein were associated with higher postdexamethasone co~isol levels. In this study the elevations of haptoglobin and alpha-l-chymotrypsin were not related to the mean afternoon cortisol level, but both were associated with a blunted TSH response to TRH. This association has not previously been reported. However, in this group of patients we do not have clear evidence of a general cortisol hypersecretion, as assessed by the mean afternoon cortisol level. It needs to be recalled though that elevated mean afternoon cortisol levels and dexamethasone suppression test nonsuppression are not necessarily identical (ttolsboer et al 1984). The interesting question that arises from this research is the nature of the association between elevations in acute phase proteins and blunting of the TSH response to TRH. On present data it is not possible to establish whether the protein changes could be a result of the hypothalamic-pituitary-thyroid axis abnormality, or vice-versa, or whether both are reflections of some other underlying abnormality in depression. In evaluating the timings from this study a number of methodological issues should be noted. All the control men underwent their neurobiological assessment on an outpatient basis, as did 20 of the 21 depressed men. However, the alcoholic men were all inpatients, and had been inpatients for about 3 weeks at the time of their neurobiological assessment. The levels of the acute phase proteins in the control men are generally about 80%, which is perhaps lower than would be expected from a control group given that the assays are standardized to a pooled control plasma giving levels of !q0%. However, a number of issues are of note in this regard. In the first instance the pooled plasma for the assay standard comprises plasma from both men and women, whereas we only measured levels in men so as to avoid any possible effects of gender. The second factor of note is that the blood samples in this study were drawn through an indwelling intravenous catheter,
1040
BIOL PSYCHIATRY 1992;32:1035-1041
P.R. Joyce et al
while the subject rested supine, and without the need for a tourniquet. This would probably lead to lower levels of plasma proteins than if they had been drawn by venipuncture with the use of a tourniquet. This would account for the lower levels of acute phase proteins in our control subjects compared to the reference population. However, it does not invalidate the differences found between our controls and depressed or alcoholic patients, as all the subjects in this study had blood drawn in a standard manner, at the same time of the day, by the same nurse, and blood samples were subsequently handled in a comparable fashion. A third methodological issue is that we have performed multiple analysis of variances and Pearson correlations on the data collected, and have not made any explicit adjustment for having performed multiple statistical tests. However, the significance of many of the findings is well beyond a 0.05 level of significance, and within the general pattern of results there is a consistency to the fir,dings that provides further support for their significance. This study could thus be taken to in part support the initial findings of Nemeroff et al (1990), which prompted this research. However, the elevated levels of alpha-l-acid glycoprotein, are probably only one aspect of a more general metabolic disturbance in plasma proteins in depressed patients. ~f alpha- l-acid glycoprotein is part of this general elevation of acute phase plasma proteihs, and is an inhibitor of plateiet 5HT/imipramine binding site,~ then this could possibly be of relevance to serotonergic mechanisms in depression. However, our findings also raise a whole series of further questions as to why such a generalized increase in a variety of acute phase plasma proteins and immunoglobulin G occurs in major depression, and what are the mechanisms that produce this increase? In considering the possible mechanisms for this generalized elevation of plasma proteins it should be recalled that the depressed patients in this study had been depressed for months or years prior to participating in this study, which also raises the question as to why such an acute phase response can presumably persist over a prolonged period of time, We wish to thank Mrs, lsobel Stevens and Robyn Abbott for their assistance in completing this study. This project was funded by grants from the Medical Research Council of New Zealand and the Alcoholic Liquor
Advisory Council,
References Abraham KI, leni JR, Meyerson LR (1987): Purification and properties of a human plasma endogenous modulator for the platelet tricyclic binding/semtonin transport complex. Biochem Biophys Acta 932:8-21. Alper CA (19"/4): Plasr,:a protein measurements as a diagnostic aid. N Engl J Med 291:287-291. Hamilton M (1960): A rating scale for depression. J Neurol Neurosurg Psychiatry 23:56-62. Healy D, Calvin J, Whitehouse AM, et al (19~i): Alpha-i-acid glycoprotein in major depression and eating disorders. J Affective Disord 22:13-20. Holsboer F, Gerken A, Steiger A, Fass V (1984): Mean 1400-1700 plasma cortisol concentration and its relationship to the Img dexamethasone suppression response in depressions and controls. Acta Psychiatr Scand 69:383-390. Kehoe WA, Kwentus JA, Sheffel WB, Harralson AF (1991): Increased alpha-l-acid glycoprotein in depression lowers free fraction of imipramine. Biol Psychiatry 29:489-493. Kushner ! (1988): The acute phase response: An overview. In Methods in Enzymology. Vol 163, pp 373-383.
Acute Phase Plasma Proteins in Depression
BIOL PSYCHIATRY 1992;32:1035-1041
1041
Laurell CB (1972): Electrophoretic and electro-immunochemical analysis of proteins. Scand J Clin Lab Invest 29:(suppl 124). Lewis JG, Elder PA (1985): An enzyme-linked immunosorbent assay (ELISA) for plasma cortisol. J Steroid Biochem 22:673-676. Maes M, Bosmans E, Suy E, Vandervorst C, Dejonckheere C, Raus J (1991): Antiphospholipid, antinuclear, Epstein-Barr and cytomegalovirus antibodies, and soluble interleukin-2 receptors in depressive patients. J Affective Disord 21:133-140. Maes M, Scharpe S, Van Grootei L, et al (1992): Higher alpha-l-antitrypsin, haptoglobin, ceruloplasmin and lower retinol binding protein plasma levels during depression: Further evidence for the existence of an inflammatory response during that illness. J Affective Disord 24:183192. Nemeroff CB, Krishnan KRR, Blazer DO, Knight DL, Benjamin D, Meyerson LR (1990): Elevated plasma concentrations of alpha-l-acid glycoprotein, a putative endogenous inhibitor of the tritiated imipramine binding site, in depressed patients. Arch Gen Psychiatry 47:337-340. Pitts AF, Carroll BT, Gehris TL, Kathol RG, Samuelson SD (1990): Elevated CSF protein in male patients with depression. Bioi Psychiatry 28:629-637.
