|Home | About | Journals | Submit | Contact Us | Français|
Neuropeptide Y (NPY) and brain-derived neurotrophic factor (BDNF) have been implicated in both the stress response and alcohol addiction. However, few studies have assessed the NPY and BDNF response to stress in alcohol-dependent participants and the concurrent measure of NPY and BDNF has not been reported in human participants.
The purpose of this study was to concurrently assess serum NPYand BDNF, as well as adrenocorticotropin (ACTH) and cortisol, in control and race- and aged-matched abstinent alcohol-dependent participants in response to a stress-inducing public-speaking task.
Basal and post-stress serum values of NPY and BDNF, as well as ACTH and cortisol, were assessed in 14 abstinent alcohol-dependent and ten healthy control male participants.
Basal measures were stable over short periods of time and stress induced a significant increase in both NPY (p= 0.002) and BDNF (p=0.006) as well as ACTH (p<0.001) and cortisol (p<0.007). Alcohol-dependent and control groups did not significantly differ on any basal or stress-induced measure. Basal and delta responses of NPY and BDNF were not significantly correlated, and delta peak responses of NPY and BDNF did not correlate with one another or with their respective ACTH and cortisol responses.
These findings reveal that both serum NPY and BDNF are responsive to behavioral stressors, although their regulatory mechanisms appear to differ from one another and those of the hypothalamic–pituitary–adrenal axis. Differences in basal and stress-induced responses of NPY and BDNF were not supported between control and abstinent alcohol-dependent subjects.
Neuropeptide Y (NPY) and brain-derived neurotrophic factor (BDNF) provide protective and neuroadaptive functions within the central nervous system (CNS). Both neuropeptides have been implicated in the modulation of stress response. NPY (or NPY mRNA) is co-released with norepinephrine (Perkins et al. 2009), is anxiolytic (Heilig et al. 1993; Yang and Gaydos 2010), and is altered following both acute (Thorsell et al. 1998) and chronic stress (Thorsell et al. 1999) in rodents. Acute stressors also elicit variable loci-dependent changes in the central expression of BDNF, although the direction and region of these changes varies across paradigms (Acuna et al. 2010; Adlard et al. 2004; Requena et al. 2010; Shi et al. 2010). Hippocampal BDNF, for example, is higher in rats exhibiting higher rates of voluntary exercise (Johnson et al. 2003) and are increased following 1 week of exercise (Adlard et al. 2004).
The role of NPY and BDNF in humans, however, has been complicated by difficulties in assessing neuropeptide functioning centrally, the variable permeability of the blood–brain barrier to neuropeptides, and the peripheral production of NPY and BDNF (Fujimura et al. 2002). Studies assessing basal or provoked measures of NPY or BDNF in humans generally support the relationship between NPY/BDNF and stress, although the data are limited. Serial cerebrospinal fluid concentrations of NPYare decreased in participants with post-traumatic stress disorder (Sah et al. 2009) and cortisol and NPY positively correlate with stress in participants immediately and 24 h following a physical and psychological stressor (Morgan et al. 2002). Intense (Rojas Vega et al. 2006) and moderate (Gold et al. 2003) exercise also increase serum BDNF in healthy humans. These studies support the relationship between NPY/BDNF and stress, although the data are limited.
Given the established connection between stress and NPY/BDNF, coupled with the role of stress in both the development and persistence of alcohol dependence, the relationship between these neuropeptides and alcohol intake has been an area of significant interest. In rodents, acute (Kinoshita et al. 2000) and chronic (Goldstone et al. 2009; Roy and Pandey 2002) alcohol intake decreases central NPYexpression. Following the cessation of chronic alcohol intake, NPY levels are altered for at least 72 h (Goldstone et al. 2009). The administration of NPY decreases alcohol reinstatement in rodents both with (Cippitelli et al. 2010) and without (Gilpin et al. 2003) exposure to a stressor. Furthermore, intrathecal NPY administration suppressed anxiety-like and arousal behaviors in rats bred for high alcohol preference (Gilpin et al. 2011). Infusion of NPY into the central amygdala produced the same effect and reduced alcohol intake (Zhang et al. 2010). Genetic variation in the NPY promoter regions also appears to affect the interplay between stress and alcohol intake. Rhesus macaques with the −1002T>G variant of NPY promoter demonstrated higher levels of arousal and increased alcohol consumption in response to a stressor (Lindell et al. 2010). Despite this provocative literature, to our knowledge serum NPY has not been assessed in alcohol-dependent subjects.
Variable loci-dependent, including striatal (McGough et al. 2004) and hippocampal (Tapia-Arancibia et al. 2001), changes in central BDNF expression also occur following chronic alcohol intake. In mice, BNDF administered into the dorsolateral striatum attenuated responding for ethanol (Jeanblanc et al. 2009) whereas the inhibition of BDNF expression increases ethanol self-administration (Jeanblanc et al. 2009). In alcohol-dependent participants, measures of serum or plasma BDNF concentrations have been mixed. Huang et al. (2008a) reported serum BNDF levels higher at 1 day relative to 7 days abstinence (although neither time point was significantly different than controls), Lee et al. (2009) found elevated plasma BDNF levels at 1 day abstinence relative to controls, and Joe et al. (2007) reported lower plasma BDNF concentrations at 30 days abstinence relative to controls. No difference in serum BDNF was found by Heberlein et al. (2010) between 1, 7, and 14 days of abstinence or between alcohol-dependent subjects and healthy controls and Umene-Nakano et al. (2009) reported similar null findings in depressed subjects with and without alcohol dependence. Both positive (Huang et al. 2008b) and negative (Heberlein et al. 2010) correlations between BDNF concentrations and alcohol withdrawal severity have been observed. Although no association between the BDNF Val66Met (G196A) polymorphism and alcohol dependence has been observed (Matsushita et al. 2004; Muschler et al. 2011; Tsai et al. 2005), this polymorphism has been related to alcohol relapse (Wojnar et al. 2009).
Thus, there exists a complex interplay between NPY/ BDNF, stress, and alcohol intake. Despite the preponderance of preclinical studies supporting a role for NPYand BDNF in the stress response, the peripheral response of these measures have received limited attention in clinical investigations. The goal in the present study was to (1) assess the response of NPY and BDNF to a psychological stressor and (2) determine if there were differences in basal and stress-induced changes in serum NPY/BDNF in abstinent alcohol-dependent participants relative to controls. We utilized the Trier Social Stress Task (Kirschbaum et al. 1993), a standardized stress protocol shown to induce subjective and physiological stress responses. Participants were a subset of a larger sample participating in a study assessing alterations in stress reactivity in alcohol-dependent participants.
Fourteen male alcohol-dependent participants were recruited from patients in residential treatment for alcohol dependence at the Dallas VA Medical Center and Homeward Bound, Inc. Patients met Diagnostic and Statistical Manual of Mental Disorders– Version IV (DSM-IV) criteria for alcohol dependence, reported alcohol as their primary drug of choice and heavy drinking for at least 90 days prior to admission, and were between 4 and 6 weeks abstinent at the time of the study. Exclusion criteria included other active mood, anxiety (except PTSD), psychotic, eating, and cognitive DSM-IV Axis I disorders not associated with alcohol use, major medical disorders or medical disorders affecting the hypothalamic–pituitary–adrenal (HPA) axis or autonomic nervous system, body mass index of ≥30, major head trauma, use of any medications that may interfere with HPA axis or autonomic nervous system functioning [e.g., psychotropics (including benzodiazepines, antidepressants, antipsychotics, mood stabilizers, anxiolytics, or beta blockers), anti-hypertensives (except hydrochlorothiazide), or hypoglycemic agents] within 2 weeks of the study.
Ten male controls reported no lifetime history of a DSM-IVAxis I substance dependence disorder (except nicotine) or an active non-substance use disorder. Other inclusion/exclusion criteria were similar to that of alcohol-dependent participants. Controls with a single first-degree or two second-degree relatives with substance dependence disorder were excluded.
After a complete description of the study was provided to the participants, consent was obtained. Participants were financially compensated for their participation. Study approval was obtained from the Institutional Review Board of the University of Texas Southwestern Medical Center at Dallas and the VA North Texas Health Care System. Participants obtained a medical exam, routine laboratory testing, electrocardiogram, and urine drug screens. Axis I disorders were assessed using the Structured Clinical Interview for DSM-IV Lifetime (First et al. 1996). Depressive symptoms were assessed with the Beck Depression Inventory (BDI) (Beck et al. 1979) and anxiety symptoms with the State Trait Anxiety Inventory (STAI) (Speilberger 1971). The Drinker Inventory of Consequences–Lifetime Consequences (Miller et al. 1995) assessed lifetime severity of alcohol-related problems. A Time-Line Follow Back (Sobell and Sobell 1978) assessed 90-day and lifetime drinking history. Alcohol-dependent participants were housed on a residential treatment unit until studies were initiated.
Participants arrived at 1700 h and were provided dinner. An intravenous catheter was inserted at 1800 h, and the Brief Symptom Inventory (BSI) (Derogatis and Melisaratos 1983) was administered 45 min later. The BSI uses 53 questions to assess cognitive, sensory, and affective changes. Nine symptoms clusters are measured as well as three global measures (Positive Symptom Distress Index (severity of response), positive symptom total (frequency of response; PST), and Global Severity Index (severity frequency of response; GSI). The task was begun at 1900 h. The Trier Social Stress Test (TSST), performed in front of a male and female confederate, involved a 5-min preparation period for a 5-min public-speaking task, followed by an arithmetic task for an additional 5 min. Participants were then asked “How stressful did you experience the experiment?” (scored 1 to 5). Basal blood samples were taken at 1845 and 1855 h followed by post-TSST samples at 1920, 1930, 1940, 1950, 2000, and 2010 h. Preliminary analyses of all time points in four individuals (two controls and two patients) revealed peak NPYand BDNF responses at 1920 and 1930 h. Due to the expense of the NPY and BNDF assays, subsequent assays were only performed for the 1845, 1855, 1920, and 1930 h samples in order to assess basal and peak concentrations. Results below include the relevant samples from the initial four subjects.
Serum BDNF and NPY were collected in redtopped vacutainer tubes and kept at room temperature for 30 min following collection (allowing coagulation). Plasma adrenocorticotropin (ACTH) and cortisol were collected in purple-topped vacutainer tubes with EDTA and were processed immediately following collection. Samples were than centrifuged (2,500, 4°C, 10 min). Samples were placed in cryovials and frozen at −80°C until assayed. NPY was measured by radioimmunoassay (ALPCO Diagnostics, Salem, NH). Calibrators were 9.4 to 300 pmol/L, the lowest detectable concentration was 3 pmol/L, and intra- and inter-assay variability was 4.5–9.2%. BDNF was measured using a sandwich ELISA kit (ChemiKine, Millipore Corp.). Calibrators were 7.8 to 500 pg/mL, the lowest detectable concentration was 7.8 pg/mL, and intra- and inter-assay variability was between 3.7% and 8.5%. ACTH was measured using an ELISA kit (Biomerica, Newport Beach, CA). Calibrators were 7 to 535 pg/mL, the lowest detectable concentration was 0.46 pg/mL, and intra- and inter-assay variability was 3.1– 6.2%. Cortisol was measured using an EIA kit (Diagnostic Systems Labs, Webster, TX). Calibrators were 0.5–60 μg/dL, the lowest detectable concentration was 0.1 μg/dL, and intraand inter-assay variability was 2.4–12%. Samples were diluted 1:50 to 1:100.
Comparisons between basal measures were assessed with paired t tests. Repeated measures analysis of variance assessed change over time (pre- to peak post-stress response) and group (control vs. alcohol dependent) interactions, followed by post hoc t tests to determine group differences in basal, peak, and delta (peak– basal) concentrations. Analyses of covariance were used to assess the potential influence of BSI scores on the BDNF and NPY comparisons. The relationship between basal and delta (Δ) peak response was assessed with Pearson's r (p<0.01). There were missing BDNF values for one control and one patient and NPY values for one patient, so these participants were not included in the respective analysis. Raw BSI scores were utilized.
Control and alcohol-dependent patients did not differ in race, employment status, and age, although alcohol-dependent participants had fewer years of education and were more likely to be smokers. The patient group also had higher BDI) and STAI (Speilberger 1971) scores compared with controls. Liver enzymes (aspartate aminotransferase and alanine aminotransferase) were not significantly different between groups. Three patients had active co-morbid substance dependence; one with cocaine dependence, one with marijuana dependence, and one with cocaine, amphetamine, and cannabis dependence. Two patients had PTSD.
Both groups found the public-speaking task equally stressful, based upon both subjective and biological measures. Both alcohol-dependent (4.0±1.2) and control (4.0±0.5) groups endorsed marked subjective stress (t=0.00, df=17.7, p=1.00). A significant time effect (mean basal vs. peak) was observed for both ACTH (F=14.34, p= 0.001) and cortisol (F=9.02, p=0.007), with no significant group×time effect for either hormone (ACTH: F=0.06, p= 0.81; cortisol: F=0.32, p=0.58) (see Fig. 1).
No significant differences were evident between the two basal measures of NPY (t=0.33, df=23, p=0.97) or BDNF (t=0.43, df=18, p=0.66), revealing short-term stability of the neuropeptides at rest. Thus, the average of both basal measures was used for subsequent analyses. Slightly more subjects showed higher net peak concentrations for NPYand BDNF at the second (1930 h), relative to the first (1920 h), post-TSST time point (NPY, 12 of 23 subjects (52%); BDNF, 14 of 22 subjects (64%).
A significant time effect (pre- vs. post-stress) was observed for both NPY (F=13.22, p=0.002) and BDNF (F=9.50, p=0.006), with no significant group (NPY: F=26, p=0.61; BDNF: F=0.18, p=0.67) or group×time (NPY: F=0.87, p=0.36; BDNF: F=0.00, p=0.99) effect for either hormone. Post hoc comparisons revealed no significant basal, peak, or Δgroup differences of serum NPYor BDNF concentrations.
Although median BSI scores of both groups tended to be low at baseline, alcohol-dependent subjects scored significantly higher on somatization, obsessive–compulsive, depression, anxiety, and psychoticism as well as the GSI and PST global scores. As BNDF and NPY have been related to depressive and/or anxiety symptoms (Duncan et al. 2009; Hashimoto 2007; Shimizu et al. 2003), the global score PST was considered as a covariate in the BNDF- and NPY-repeated measures and post hoc t test comparisons. (PST was chosen as it demonstrated the greatest difference between groups. Only one covariate was assessed due to the relatively small sample size). The time effect for NPY no longer reached significance (F=3.41, p=0.08); the time effect for BNDF persisted (F=4.63, p=0.046). No significant group differences over time were revealed. No significant group differences became evident in the post hoc comparisons (baseline, peak, and Δpeak) when PST was considered as a covariate. In addition, correlations between baseline, peak, and Δpeak with PST did not yield significant relationships in the combined sample.
There were no significant relationships between basal and delta peak concentrations of NPY, BDNF, ACTH, or cortisol in either group. Within-group Δpeak responses were significantly correlated for ACTH and cortisol (control: r=0.92, n=10, p<0.001; alcohol dependent: r=0.98, n=13, p<0.001), but no other neuropeptide/hormone peak response relationships were evident. Since almost all (83%) participants endorsed high subjective stress (four or five), an association between subjective stress and neuropeptides increase was not assessed.
To our knowledge, this is the first published study to assess (1) the concurrent NPY and BDNF (as well as ACTH and cortisol) response to stress, (2) two consecutive basal measures of NPY and BNDF, (3) NPY in alcohol-dependent participants, and (4) the NPY and BDNF responses to stress in alcohol-dependent participants. The findings demonstrate that both serum NPY and BDNF concentrations significantly increase in response to an acute psychosocial stressor, although these changes were somewhat mitigated by baseline mood and anxiety symptoms. Similar basal measures suggest relative stability of NPYand BDNF over short periods of time. Neither basal nor peak measures of NPYor BDNF differed between control and 4– 6-week abstinent alcohol-dependent participants, suggesting either a resistance to alcohol-induced modulation or a normalization by 4 weeks abstinence.
The NPY and BDNF responses to stress did not significantly correlate with each other or with the ACTH or cortisol response. These findings may suggest different sources, secretory dynamics, and/or regulatory mechanisms for NPY, BDNF, and pituitary–adrenal responses. NPY is co-released with norepinephrine from peripheral sympathetic neurons (Perkins et al. 2009), so the increase following stress is not unexpected. Other peripheral sources of NPY include platelets (Myers et al. 1993) and adipocytes (Yang et al. 2008), although their relevance to stress-induced NPY is unknown. Increases in plasma NPY following US Navy survival school (Morgan et al. 2002) and public speaking also suggest NPY responsivity to variable levels of stress. Peripheral sources of BDNF include platelets (Fujimura et al. 2002), peripheral blood mononuclear cells (Lee and Kim 2010), and vascular cells (Scarisbrick et al. 1993), although the mechanism of BDNF release during stress has not been identified. Our present findings, coupled with those observed following exercise (Gold et al. 2003; Rojas Vega et al. 2006), suggest that BDNF increases in response to both physical and psychological stressors and can occur with in 20-min post-stress. Although the time required to release BDNF during a stressor is unknown, our findings suggest that 20 min post-stress initiation is sufficient for BDNF to be released into the vasculature and taken up into the platelets. Presumably, BDNF is subsequently released into the serum from platelets during the coagulation process (Karege et al. 2005).
Our null group basal BDNF findings are similar to those of others (Heberlein et al. 2010; Huang et al. 2008b; Umene-Nakano et al. 2009). Although the higher BDNF concentrations reported by Lee et al. (2009) in alcohol-dependent subjects may have been due to the presence of withdrawal, Joe et al. (2007) found lower BDNF levels in a 30-day abstinent, healthy population of alcohol-dependent subjects similar to ours. Possible explanations for these differences include the use of fasting, morning plasma measures by Joe et al. (2007) compared with the present sample of post-prandial, evening serum levels, particularly as BDNF has been shown to exhibit a diurnal cycle similar to cortisol (Begliuomini et al. 2008). In addition, reports of basal BDNF concentrations have fluctuated markedly between laboratories (Heberlein et al. 2010; Huang et al. 2008b; Joe et al. 2007; Umene-Nakano et al. 2009), suggesting differences in assay techniques. The serum BDNF concentrations reported in this manuscript are within the range usually observed in previous studies (Karege et al. 2005).
Unexpectedly, our alcohol-dependent population did not demonstrate a muted ACTH or glucocorticoid response to the behavioral stressor relative to controls, as has been reported following public-speaking stressors (Junghanns et al. 2003; King et al. 2006; Lovallo et al. 2000), exercise (Coiro et al. 2007), cold pressor and mental arithmetic (Errico et al. 1993), and isometric handgrip and public speaking (Bernardy et al. 2003). Other studies have also revealed a blunted ACTH and/or cortisol response in alcohol-dependent subjects administered various pharmacological stimuli (Adinoff et al. 2005; Coiro and Vescovi 1999; Wand and Dobs 1991). There is no straightforward explanation for the present finding. It is possible that the NPY and BDNF responses would have shown group differences if measured in a patient group demonstrating subdued ACTH and/or cortisol stress responses.
Strengths of the present study included concurrent measures of BDNF and NPY during a psychosocial stressor, documented by both subjective stress and ACTH/ cortisol measures. Controls were similar in age and race to the alcohol-dependent participants. Alcohol-dependent participants were post-withdrawal and not taking medications that interfered with CNS, HPA axis, or autonomic nervous system functioning. However, serum measures may have missed group differences that occurred following peak effects and the period of abstinence may have missed changes present earlier in the withdrawal process. Although the sample size was modest, the findings were relatively unequivocal. Although several studies have noted a relationship between nicotine use/administration and BDNF and NPY concentrations (Bhang et al. 2010; Chen et al. 2007; Frankish et al. 1995; Kim et al. 2007; Li et al. 2000), the unequal distribution of nicotine dependence between the control and alcohol-dependent groups would have tended to result in a Type I, rather than Type II, error. The alcohol-dependent subjects also endorsed higher mood and anxiety symptoms relative to controls, which appeared to contribute to the NPY and BDNF response. Thus, subjects with greater mood and anxiety symptoms may show a heightened NPY and BNDF stress response. As only men were assessed, these findings may not be applicable to women.
These findings support the release of both peripheral NPYand BDNF during psychological stress yet suggest that distinct mechanisms underlie their release. Both peripheral NPY and BDNF are widely expressed, although their presence in certain tissues serve unknown functions and act via unclear mechanisms. Future research should be directed toward examining the physiological role of both neuropeptides in mediating their unique contributions to the acute stress response.
The authors thank Homeward Bound, Inc. and the Dallas VA Substance Abuse Team for their assistance in the recruitment and clinical care of patients and the UT Southwestern CTRC staff for their excellent patient care and meticulous attention to research protocol.
Role of funding source Funding for this study was provided by NIH INIAStress U01AA13515, Department of Veterans Affairs, and NIH CTSA grant UL1 RR024982. None of the funding sources had a role in study design; in the collection, analysis, and interpretation of data; or in the writing of the report.
Conflict of interest CN discloses employment by VA North Texas Health Care System, Dallas, TX, USA. Points of view in this document are those of the author(s) and do not necessarily represent the official position of NIMH, the Department of Veterans Affairs, or the US Government. Dr. North also discloses research support from NIAAA, NIMH, NIDDK, the American Orthopedic Association, the American Psychiatric Association, the Department of Veterans Affairs, and UT Southwestern Medical Center; consulting fees from the University of Oklahoma Health Sciences Center, Cubic, Inc., and the National Center for PTSD (Research Education in Disaster Mental Health); and honoraria from Washington University in St. Louis, the University of Alabama at Tuscaloosa, the Assisi Foundation of Memphis, and Magellan Health Services. BA received grant support from NIAAA, NIDA, and the Department of Veterans Affairs; consulted for Teva Pharmaceutical Industries Ltd., Shook, Hardy and Bacon LLP (medical malpractice for tobacco companies), and Paul J. Passante, P.C. (medical malpractice), and received honoraria from the Medical University of South Carolina, American Institute of Biological Sciences, American Academy of Addiction Psychiatry, Methodist Medical Center (Dallas, TX), Vanderbilt University, John Peter Smith Hospital, University of North Texas Health Science Center; American Academy of Addiction Psychiatry, University of New Mexico. YM, TW, UR, HX, and MJ declare that they have no conflicts of interest.
Donna Meng, Department of Psychiatry, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-8564, USA.
TingChin Wu, Department of Psychiatry, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-8564, USA.
Uma Rao, Center for Molecular and Behavioral Neuroscience and Department of Psychiatry and Behavioral Sciences Meharry Medical College, 1005 Dr. D.B. Todd Jr. Boulevard, Nashville, TN 37208, USA.
Carol S. North, Department of Psychiatry, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-8564, USA. VA North Texas Health Care System, 4500 S. Lancaster Rd, Dallas, TX 75216, USA.
Hong Xiao, Department of Family and Community Medicine, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-8564, USA.
Martin A. Javors, Department of Psychiatry, University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, TX 78229, USA.
Bryon Adinoff, Department of Psychiatry, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-8564, USA. VA North Texas Health Care System, 4500 S. Lancaster Rd, Dallas, TX 75216, USA.