Changes in neuroactive steroid secretion associated with CO2-induced panic

attacks in normal individuals

a * b c d

Francesca Brambilla ,Giulia Perini, , Mariangela Serra , Maria Giuseppina Pisu, , Stefano b b e f cZanone, , Tommaso Toffanin , Stefano Milleri , Cristina Segura Garcia , Giovanni Biggio

a Department of Psychiatry, Sacco Hospital, Milano 20157, Italy

b Department of Neuroscience, Padova University, Padova 35128, Italyc Department of Life and Environmental Sciences, University of Cagliari, Cagliari 09100, Italyd CNR Institute for Neurosciences, Cagliari 09100, Italye Centro Ricerche Cliniche, G.B.Rossi Hospital, Verona 37134, Italy

f Department of Psychiatry, Catanzaro University,Catanzaro, Italy

Running title: Neuroactive steroids and CO2-induced panic attacks in normal individuals

Corresponding author:

*dr. Prof. Francesca Brambilla

Centro di Psiconeuroendocrinologia,

Piazza Grandi 3

Milano 20129

Italy

Tel.: +39-02-717350 or +39-368-3017420. Fax: +39-02-70122889.

E-mail: francesca.brambilla4@tin.it

1

Summary

Neuroactive steroids modulate anxiety in experimental animals and possibly in humans. The

secretion of these compounds has been found to be altered in Panic Disorder (PD), with such

alterations having been suggested to be a possible cause or effect of panic symptomatology.

Paniclike attacks can be induced in healthy individuals by administration of panicogenic agents or

by physical procedures, and we have now measured the plasma concentrations of neuroactive

steroids in such individuals before, during, and after panicogenic inhalation of CO2 in order to

investigate whether abnormalities of neuroactive steroid secretion might contribute to the

pathogenesis of PD. Fifty-nine psychologically and physically healthy subjects, including 42

women (11 in the follicular phase of the menstrual cycle, 14 in the luteal phase, and 17 taking

contraceptive pills) and 17 men, who experienced a paniclike attack on previous exposure to 7%

CO2 were again administered 7% CO2 for 20 min. Thirty-three of these individuals (responders)

again experienced a paniclike attack, whereas the remaining 26 subjects did not (nonresponders).

All subjects were examined with the VAS-A and PSL-III-R scales for anxiety and panic

symptomatology before and after CO2 inhalation. The plasma concentrations of progesterone,

3α,5α-tetrahydroprogesterone (3α,5α-THPROG = allopregnanolone), 3α,5α-

tetrahydrodesoxycorticosterone (3α,5α-THDOC), dehydroepiandrosterone (DHEA), and cortisol

were measured 15 min and immediately before the onset of CO2 administration as well as

immediately, 10 min, 30 min, and 50 min after the end of CO2 inhalation. Neuroactive steroids were

measured in the laboratory of prof. Biggio in Cagliari, Sardinia, Italy. Neurosteroid levels did not

change significantly in both responders and nonresponders before, during, or after CO2 inhalation.

These data suggest that neuroactive steroid concentrations before, during, or after CO2 inhalation do

not seem to correlate with panic symptomatology during paniclike attacks in subjects not affected

by PD, and they therefore do not support the notion that abnormalities in neuroactive steroid

secretion are either a cause or an effect of such attacks.

Keywords: CO2 inhalation, panic attack, progesterone, 3α,5α-tetrahydroprogesterone,

3α,5α-tetrahydrodesoxycorticosterone, dehydroepiandrosterone, cortisol.

2

1. Introduction

Neurosteroids produced in the central nervous system as well as centrally acting neuroactive

steroids secreted in the periphery modulate the expression of anxiety in experimental animals

(Beaulieu, 1998). These compounds regulate neuronal excitability with rapid, nongenomic effects

that are initiated at the cell surface through agonistic or antagonistic interaction with γ-aminobutyric

acid type A (GABAA) receptors inserted in the membrane of glutamatergic neurons of the amygdala

and hippocampus (granular cells, pyramidal and pyramidal-like neurons) acting by intracellular

membrane diffusion after being produced by these neurons (Wieland et al., 1991; Paul and Purdy,

1992; Patchev et al., 1994; Barbaccia et al., 1996; Brot et al., 1997; Beaulieu, 1998; Concas et al.,

1998; Akwa et al., 1999; Rupprecht and Holsboer, 1999; Bitran et al., 2000; Engel and Grant, 2001,

Akk et al., 2005, Agis Balboa et al., 2006, Agis Balboa et al,, 2007, Pinna et al. 2008, Gartside et al.

2010 ).

Although neuroactive steroids have been implicated in humans in the response to stress and in

certain neurological and psychiatric conditions, such as depression, ethanol withdrawal, and

epilepsy, data obtained from individuals with anxiety disorders, including panic disorder (PD), have

been inconclusive with regard to a potential pathogenic role for these compounds (Rupprecht and

Holsboer, 1999; Barbaccia et al., 2000; Bicikova et al., 2000; Spivak et al., 2000; Semeniuk et al.,

2001; Heydary and Le Melledo, 2002, Rupprecht 2003). Either normal or greater than normal levels

of anxiolytic neuroactive steroids, which act as agonists at GABAA receptors, or reduced levels of

neuroactive steroids that act as GABAA receptor antagonists have been detected in men or women

with PD in the periods between panic attacks, whereas no changes in neuroactive steroid secretion

were observed in such individuals after clinically successful short- or longterm pharmacological

therapy (Bicikova et al., 2000; Strohle et al., 2002; Brambilla et al., 2003; Brambilla et al., 2004;

Pisu and Serra, 2004; Brambilla et al., 2005; Eser et al., 2005). Panic attacks induced in PD patients

pharmacologically with pentagastrin, sodium lactate, or cholecystokinin tetrapeptide have been

found to be accompanied by no change in neuroactive steroid levels, or by a reduction in the level

of neuroactive steroids that act as GABAA receptor agonists, or an increase in that of those that act

as receptor antagonists (Zwanzger et al., 2001; Tait et al., 2002; Strohle et al., 2003; Zwanzger et

al., 2003; Eser et al., 2005; Dell’Osso et al., 2009), suggesting that such changes might contribute to

panic symptomatology in PD patients. On the other hand, individuals with subthreshold panic-

agoraphobic symptomatic spectrum have elevated blood levels of the anxiolytic GABAA receptor

agonist dehydroepiandrosterone (Dell’Osso et al., 2009). Despite data suggestive of a role for

3

neurosteroid secretion in the pathogenesis of PD, it remains unclear whether altered secretion

between or during panic attacks is a cause or consequence of the psychopathology, in the last case

possibly acting as a buffer for the anxiety disorder (Rupprecht et al., 2001),

Panic attacks can be induced in apparently normal individuals by acute stimulation with

panicogenic substances or by physical panicogenic procedures, possibly as a result of a preexisting

elevated anxiety or “anxiety sensitivity” or of specific personality traits such as physical

aggressiveness, irritability, somatic anxiety, or stress susceptibility (Perna et al., 1995; Zinbarg et

al., 2001; Perna et al., 2003; Coryell, 2004; Battaglia et al., 2007; Schmidt and Zvolenski, 2007;

Battaglia et al., 2009; Esquivel et al., 2009; Toru et al., 2010). Increased secretion of 3α,5α-

tetrahydrodesoxycorticosterone (3α,5α-THDOC), adrenocorticotropic hormone, and cortisol was

detected in healthy volunteers in association with cholecystokinin tetrapeptide–induced panic

attacks, and was suggested to reflect activation of the hypothalamic-pituitary-adrenal axis

contributing to termination of the anxiety-stress response through enhancement of GABAA receptor

function (Eser et al., 2005). Given that some apparently normal individuals might develop PD in

response to specific life experiences or to stimulation with panicogenic substances, we set out to

study the secretory patterns of neuroactive steroids in healthy subjects who developped panic

attacks after a previous CO2 inhalation, in an attempt to determine the possible influence of

neurosteroid alterations on predisposition to PD. Our subjects could be considered as predisposed

to PD, and therefore to be intermediate between normal subjects and panic patients. Inhalation of

CO2 affects neurosteroid secretion in experimental animals (Barbaccia et al., 1996), and it induces

panic attacks in both normal individuals and PD patients when administered at 7% over 20 min or at

35% in a single breath (Gorman et al., 1990; Perna et al., 1995; Coryell, 2004; Battaglia et al.,

2007). We administered air containing 7% CO2 for 20 min to a group of physically and

psychologically healthy subjects with no personal or familial history of psychopathologies of any

type. Consistent with previous observations, some of the subjects (responders) experienced acute

panic symptomatology that was immediately extinguished by interruption of CO2 inhalation. After

a period of 1 to 11 months, during which the subjects did not experience spontaneous panic attacks,

we again administered 7% CO2 for 20 min to the responders. These subjects were perfectly

conscious that the eventual panic attack could be immediately blocked by interrupting CO2

inhalation , were absolutely not scared by the procedure which had been amply explained to them,

accepted the risk of a panic attack knowing by the precedent experience that it was innocuous and

was not going to occur spontaneously thereafter. Only a subset of these individuals experienced

another panic crisis, which was not more severe than the first one but was severe enough to make

4

some of them to interrupt the inhalation, and we measured the neuroactive steroid response to the

CO2 stimulus in these responders and nonresponders.

The aim of our study was to determine whether neuroactive steroid secretion before, during, or after

CO2 inhalation might differ between the psychiatrically healthy subjects who developed

(responders) and those who did not developed (nonresponders) panic attacks during CO2 inhalation,

the former possibly representing candidates for full-blown Panic Disorder, their neuroactive steroid

impairments being a putative biological background for the disease.

2. Methods

2.1. Recruitment and assessment of study subjects

Fifty-nine individuals, 42 women and 17 men, who had previously responded with a typical panic

attack to inhalation of 7% CO2 for 20 min were entered into the study. All subjects provided written

informed consent to participation in the study, which was approved by the appropriate institutional

ethics committees. They were physically healthy as revealed by physical examination and routine

laboratory tests, were aged 19 to 47 years (mean ± SD, 33 ± 14 years), and had no current or history

of psychopathologies of any type according to DSM-IV criteria (American Psychiatric Association,

1994) and as demonstrated in a clinical interview with an expert psychiatrist using the M.I.N.I.

5.0.0 scale (Sheehan et al., 1994). Exclusion criteria for the study included current medical

disorders (in particular, cardiocirculatory or respiratory disorders), epilepsy or other neurological

disorders, immunopathies, allergopathies, endocrinopathies, obesity, or recent weight loss, current

pregnancy, personal or familial history of aneurism, present or past alcohol or drug abuse, current

use of medical or psychiatric drugs of any type, and psychotropic abnormalities based on DSM-IV

criteria. The female subjects experienced regular menstruation every 28 to 30 days; 17 of them were

on low-dose combined estrogen-progestin contraceptive pills..

The probands were divided in 4 groups , women in the follicular phase of the cycle, women in the

luteal phase of the cycle, women on contraceptive pills and men, because they obviously secrete

different amounts of progesterone and progesterone derivatives.

2.2.Design and procedure of the study

The CO2 test was initiated between 08:30 and 09:00 hours. We selected inhalation of 7% CO2 for

20 min rather than that of 35% CO2 in one breath because the longer time required for

administration of the gas and the induction of panic-like attacks in the 7% procedure would better

5

allow evaluation of the panic symptomatology by both subjects and physicians, the blood drawing

and the administration of psychological rating scales, while the two procedures stimulate the same

panic attack. Subjects came to the laboratory after a 12-h fast and rested in a chair for 30 min, after

which an edetic acid–anticoagulant cannula was inserted into a forearm vein for blood collection.

For the test, the subjects breathed through a mask connected to two balloons by tubing and a three-

way valve. Each balloon was connected to a cylinder containing either compressed air or air

enriched with 7% CO2. A sensory apparatus (Jaeger Oxycon Champion) for measurement of

respiratory parameters was inserted between the mask and the tube. Each subject could ask to

interrupt the CO2 inhalation if the panic symptomatology became too severe. The Visual Analogue

Scale for Anxiety (VAS-A: range from 0 = no anxiety to 100 = maximum anxiety) (Woods et al.,

1987; Battaglia and Perna, 1995) and the Panic Symptom List III Revised (PSL-III-R) derived from

the Panic Attack Definition in DSM-III-R (Pols et al., 1991; Battaglia and Perna, 1995) were

administered both immediately before (T0) and after (T20) CO2 administration. The PSL-III-R scale

consists of 13 items, which are the most typical for panic attacks, including sensation of shortness

of breath or suffocation; feeling dizzy, unsteady, or faint; palpitation or accelerated heart rate;

trembling; sweating; feeling of choking; nausea or abdominal distress; sense of unreality; numbness

or tingling sensation; hot flushes or chills; chest pain or discomfort; fear of dying; and fear of losing

control or going crazy. We used the PSL III R scale because it was always used by us in previous

work on PD and because it takes into consideration a full list of symptoms specific for Panic

Disorder. Subjects were then classified as responders or nonresponders with regard to the induction

of paniclike symptomatology by CO2 inhalation on the basis of VAS-A and PSL-III-R scores.

Responders were defined as subjects who showed an increase in VAS-A score [(VAS-A posttest) –

(VAS-A pretest)/100 – (VAS-A pretest)]of >26% and an increase of >4 in the total PSL-III-R score

(with an increase of at least 1 for at least 4 of the 13 symptoms).

Blood samples for measurement of progesterone, 3α,5α-tetrahydroprogesterone

(allopregnanolone, 3α,5α-THPROG), 3α,5α-tetrahydrodesossicorticosterone (3α,5α-THDOC),

dehydroepiandrosterone (DHEA), and cortisol were collected 15 min (T–15) and immediately (T0)

before the onset as well as at the end of CO2 administration (T20), then 10 min (T30), 30 min (T50),

and 50 min (T70) after the end of CO2 inhalation. The blood was immediately centrifuged, and the

isolated plasma was stored at –80°C until assayed. Among the 25 female subjects not taking

contraceptive pills, the test was performed during early follicular phase of the menstrual cycle (days

7 to 8 of the cycle) in 11 women and during middle luteal phase (days 22 to 23 of the cycle) in 14

women. The probands were divided in the 4 groups (women in the follicular phase of the cycle,

6

women in the luteal phase of the cycle, women on contraceptive pills, men) because they

obviouslysecrete different amounts of progesterone and progesterone derivatives.

2.3. Steroid assays

The neuroactive steroids were measured in the laboratory of prof. Biggio,Cagliari, Sardegna,Italy.

EDTA-anticoagulated blood was drawn, immediately centrifuged and the plasma frozed at -80° C

until assayed. All plasma samples were analyzed together in order to avoid possible consequences

of differences in reagent batches. Plasma (1 ml) was diluted with 2 ml of water and then subjected

to extraction three times with 3-ml batches of ethyl acetate. The recovery (90%) of steroids through

the extraction procedure was monitored by the addition of a trace amount of tritiated cortisol (6000

to 8000 cpm, 52 Ci/mmol; New England Nuclear). Steroids were quantified by radioimmunoassay

as previously described (Serra et al., 1999). The specific antibodies to progesterone, to DHEA, and

to cortisol were obtained from ICN (Costa Mesa, CA); those to 3α,5α-THPROG and to 3α,5α-

THDOC were generated in sheep and rabbits, respectively. The specificity of the antibodies to

3α,5α-THPROG was characterized previously (Purdy et al., 1991); they showed no cross-reactivity

with other steroids including.

The limit of detection for the radioimmunoassays, expressed as the minimal amount of steroid

distinguishable from the blank sample, was 0.01 ng. Intra- and interassay coefficients of variation

for each steroid ranged between 5 and 7% and between 9 and 11% respectively.

2.4. Statistical analysis

Data were analyzed statistically using the SPSS 18 (IBM Corporation, Armonk, N.Y., USA) for

Windows. Data are presented as means, standard deviations and frequency of occurrence (%).

Firstly, a transformation into natural logaritms of the data regarding the concentrations of

progesterone, allopreganolone, 3α,5α-tetrahydrodesoxycorticosterone, dehydroepiandrosterone and

cortisol was made. Then a first Analysis of Variance (ANOVA) for all variables with the exception

of progesterone was performed. No lack of homogeneity was found in the data. In relation to

progesterone, considering the different physiological levels among genders, an ANOVA was made

controlling for genders (females: groups 1, 2 and 4; males: group 3) and no inhomogeneity

emerged.

Differences in hormone level curves across the time between the four groups were tested by

General Linear Model (GLM) analysis for repeated measures using the response to CO2 test

7

(positive = 1; negative = 0) as covariate. Post Hoc comparisons through the Bonferroni test were

made if significant p values were found between groups comparisons.

Values of p< 0.005 were considered statistically significant.

3. Results

3.1. Clinical effects

Six of the 11 women studied during the follicular phase of the menstrual cycle, 9 of the 14

women in the luteal phase of the cycle, 10 of the 17 women on contraceptive pills, and 8 of the 17

men developed paniclike symptomatology (responders) during CO2 inhalation, with the symptoms

receding immediately after its termination. The other individuals showed no subjective or objective

pathological symptomatology of any type before, during, or after CO2 inhalation (nonresponders).

3.2. Biological and psychological effects

Table 1 shows the results of GLM for each hormone. As expected (considering the different

physiological levels among genders and the phase of the menstrual cycle among women) the only

significant difference between groups was found with regards to progesterone levels; in this case

women in the luteal phase always showed higher levels than all the other groups (Fig. 1 and 2-A).

For all the other hormones (Fig. 1and 2. B-C-D-E), no significant differences were found between

the groups at GLM and no interaction was found between the groups and the outcome (panic-like

response to CO2 test).

Table 2 shows the results of paired sample t-Test of VAS-A and PSL-III-R scores before (T0) and

after CO2 inhalation (T1). Significant differences between the two times were found in all items.

One way ANOVA did not show significant differences either in VAS-A or PSL-III-R scores

between the four groups at T1.

8

4. Discussion

Although exploratory in nature and performed with a relatively small number of subjects, our study

may shed light on the neurobiological background of panic attacks. In particular, our results suggest

that neuroactive steroid secretion under basal conditions does not differ between normal subjects

who develop paniclike symptomatology during CO2 inhalation and those who do not. Different

(panicogenic versus antipanicogenic) neuroactive steroid backgrounds may thus not delineate

responders to CO2 inhalation from nonresponders, and the paniclike attacks artificially induced in

the subjects of our study, and possibly the spontaneous attacks in PD patients, should therefore not

necessarily be ascribed to such a pathological background. The changes in neuroactive steroid

levels previously observed in PD patients during ictal and interictal phases (Bicikova et al., 2000;

Strohle et al., 2002, 2003; Brambilla et al., 2003; Brambilla et al., 2004; Pisu and Serra, 2004;

Brambilla et al., 2005) might be a result and not a cause of the disease, possibly serving as a buffer

for the panic symptomatology.

No clear-cut changes in steroid concentrations or differences between responders and

nonresponders sufficiently large to suggest an active role for neuroactive steroids in the promotion

of or protection from paniclike symptomatology were apparent in any of the four groups of subjects.

We therefore suggest that neuroactive steroid secretion during or after CO2 inhalation is not directly

related to the development of panic symptomatology in responders or to its absence in

nonresponders.

Given that the paniclike symptomatology observed in the subjects of the present study was

virtually identical in terms of type and severity to that of PD patients during a spontaneous attack,

the lack of differences in neuroactive steroid concentrations between responders and nonresponders

either before, during, or after CO2 inhalation suggests that the changes in such concentrations

reported previously in the literature in association with spontaneous or provoked panic attacks

might not be specifically linked to the attacks. However, it should be noted that, whereas cortisol

9

secretion has previously been found to remain unchanged during artificially provoked panic attacks

in PD patients, it was shown to increase markedly during spontaneous attacks in some but not all

studies (Cameron et al., 1987; Levin et al., 1987; Woods et al., 1987; Gorman et al., 1989;

Hollander et al., 1989; Seier et al., 1997; Bandelow et al., 2000; Van Duinen et al., 2004; Petrowski

et al., 2010). The reason for this difference has remained unclear, but it might be related to the lack

of “novelty” in the provoked attacks compared with the spontaneous attacks (Ursin et al., 1978;

Levine, 2000). In other words, PD patients, like normal subjects, might consider the dangerousness

of an attack to be limited when it is provoked, with the procedure being explained and the attack

occurring in the presence of a physician ready to intervene. Alternatively, in some cases, given that

PD patients are familiar with the fact that spontaneous attacks are not physically dangerous, the lack

of “novelty” might block the possible hypothalamic-pituitary-adrenal or neuroactive steroid

responses to the apparently stressful situation, consistent with the notion of “repeated hits” in the

allostatic load model described by McEwen (1998). This scenario might similarly explain the lack

of changes in neuroactive steroid secretion before and during the paniclike attacks in the subjects of

the present study. If this is the case, our results may not be directly relevant to spontaneous attacks

in PD patients. Indeed, altered neuroactive steroid secretion was apparent in PD patients both during

the interictal phase and after artificially provoked attacks (Bicikova et al., 2000; Strohle et al., 2002;

Brambilla et al., 2003; Brambilla et al., 2004; Pisu and Serra, 2004; Brambilla et al., 2005),

suggesting that it is not the attack itself as a stressful physical phenomenon that is responsible for

the changes in neuroactive steroid concentrations but possibly the psychopathology behind it.

Alternatively, as mentioned above, the changes in neuroactive steroid secretion observed in patients

with a long history of panic attacks may serve as a buffer response to the panic symptomatology, a

phenomenon not present in the normal subjects of the present study.

Whereas VAS-A and PSL-III-R scores differed significantly between responders and

nonresponders in the present study, they did not consistently correlate with neuroactive steroid

concentrations before, during, or after CO2 inhalation, This finding might be due again to the

10

relatively small number of subjects as well as to the large number of symptoms taken into

consideration.

In conclusion, our study does not appear to support the existence of a pathogenetic link

between changes in neuroactive steroid secretion and artificially induced panic attacks, or between

such changes and specific panic symptomatology. Rather, our findings suggest that alterations in

neuroactive steroid production may reflect a nonspecific response to the emotional changes

associated with panic symptomatology.

11

References

Agis-Balboa, R.C., Pinna, G., Zhubi, A., Maloku, E., Veldic, M., Costa, E., Guidotti, A. 2006.

Characterization of brain neurons that express enzyme mediating neurosteroid biosynthesis.

Proc. Natl. Acad, Sci. U.S.A. 103, 14602-14607

Agis-Balboa, R.C., Pinna, G., Pibiri,F., Kadriu, B., Costa, E., Guidotti, A. 2007. Down-regulation

of neurosteroid biosynthesis in corticolimbic circuits mediates social isolation-induced

behaviour in mice. Proc. Natl. Acad. Sci. U.S.A. 104, 18736-18741

Akk, G., Shu, H.J., Wang, C., Steinbach, J.H., Korumski, C.F., Covey, D.F., Mennerick, S. 2005.

Neurosteroid access to the GABA(A) receptor. J. Neurosci. 25, 11605-11613

Akwa, Y., Purdy, R.H., Koch, G.F., Britton, K.T., 1999. The amygdala mediates the anxiolytic

effect of the neurosteroid allopregnanolone in rat. Behav. Brain Res. 106, 119-125.

American Psychiatric Association, 1994. DSM-IV, Diagnostic and Statistical Manual of Mental

Disorders, IV ed. rev., American Psychiatric Press, Washington, DC.

Bandelow, B., Wedekind, D., Pauls, J., Brooks, A., Hajak, G., Ruther, E., 2000. Salivary cortisol in

panic attacks. Am. J. Psychiatry. 157, 456.

Barbaccia, M., Rossetti, G., Trabucchi, M., Mostallino, M.C., Concas, A., Purdy, R.H., Biggio. G,

1996. Time-dependent changes in the rat brain neuroactive steroid concentrations and

GABAA receptor function after acute stress. Neuroendocrinology. 63, 166.

Barbaccia, M.L., Lello, S., Sidiropoulou, T., Cocco, T., Sorge, S.P., Cocchiarale, A., Piermarini, V.,

Sabato A.F., Trabucchi, M., Romanici, C., 2000. Plasma 5α-androstane-3α,17βdiol, an

endogenous steroid that positively modulates GABA receptor function and anxiety: a study

in menopausal women. Psychoneuroendocrinology. 25, 659-675.

Battaglia, M., Perna, G., 1995. The 35% CO2 challenge in panic disorder: optimization by receiver

operating characteristic (ROC) analysis. J. Psychiatry Res. 29, 111-119.

Battaglia, M., Ogliari, A., Harris, J., Spatola, C.A.M., Pesenti-Gritti, P., Reichborn-Kjennerud. T.,

Torgersen, S., Kringlen, E., Tambs, K., 2007. A genetic study of the acute anxious response

to carbon dioxide stimulation in man. J. Psychiatry Res. 41, 906-917.

Battaglia, M., Pesenti-Gritti, P., Medland, S.E., Ogliari, A., Tambs, K., Spatola, C.A.M., 2009. A

genetically informed study of the association between childhood separation anxiety,

sensitivity to CO2, panic disorder, and the effect of childhood parental loss. Arch. Gen.

Psychiatry 66, 64-71.

Beaulieu, E.E., 1998. Neurosteroids: a novel function of the brain. Psychoneuroendocrinology. 23,

114-119.

12

Bicikova, M., Tallova, J., Hill, M., Rausova, Z., Hampl, R., 2000. Serum concentrations of some

neuroactive steroids in women suffering from mixed anxiety-depressive disorder.

Neurochem. Res. 12, 1623-1627.

Bitran, D., Klibanski, D.A., Martin, G.A., 2000. The neurosteroid pregnanolone prevents the

anxiogenic-like effect of inescapable shock in the rat. Psychopharmacology (Berlin) 151,

31-37.

Brambilla, F., Biggio, G., Pisu, M.G., Bellodi, L., Perna, G., Bogdanovich-Djukic, V., Purdi, R.H.,

Serra, M., 2003. Neurosteroid secretion in panic disorder. Psychiatry Res. 118, 107-116.

Brambilla, F., Biggio, G., Pisu, M.G., Purdy, R.H., Gerra, G., Zaimovich, A., Serra, M., 2004.

Plasma concentrations of anxiolytic neurosteroids in men with normal anxiety scores: a

correlation analysis. Neuropsychobiology. 50, 6-9.

Brambilla, F., Mellado, C., Alciati, A. Pisu, M.G., Purdy, R.H., Zanone, S., Perini, G., Serra, M.R.,

Biggio, G., 2005. Plasma concentrations of anxiolytic neuroactive steroids in men with

panic disorder. Psychiatry Res. 135, 185-190.

Brot, M.D., Akwa, Y., Purdy, R.H., Koob, G.F., Britten, K.T., 1997. The anxiolytic-like effects of

the neurosteroid allopregnanolone. Interactions with GABAA receptors. Eur. J. Pharmacol.

325, 1-7.

Cameron, O.G., Lee, M.A., Curtis G.C., McCann, D.S., 1987. Endocrine and physiological changes

during “spontaneous” panic attacks. Psychoneuroendocrinology. 12, 321-331.

Concas, A., Mostallino, P.C., Porcu, P., Follesa, P., Barbaccia, M., Trabucchi, M., Purdy, R.,

Grisenti P., Biggio, G., 1998. Role of brain allopregnanolone in the plasticity of γ-

aminobutyric acid type A receptor in rat brain during pregnancy and after delivery. Proc.

Natl. Acad. Sci. USA 95, 13284-13289.

Coryell, W.H., 2004. Anxiety responses to CO2 inhalation in subjects at high risk for panic

disorder. Biol. Psychiatry 55: 1S-1225.

Dell’Osso, I., Carmassi, C., Da Pozzo, E., Mundo, F., Martini, C., 2009. Correlations between

cortisol and DHEAS levels and subthreshold panic-agoraphobic spectrum symptoms in

healthy subjects. World Psychiatry 8, 51.

Engel, S.R., Grant, K.A., 2001. Neurosteroids and behaviour. Int. Rev. Neurobiology. 46, 321-348.

Eser, D., Di Michele, F., Zwanger, P., Pasini, A., Baghai, T.C., Schule, C., Rupprecht, R., Romeo,

E., 2005. Panic induction with cholecystokinin-tetrapeptide (CCK-4) increases plasma

concentrations of the neuroactive steroid 3α-5α-tetrahydrodeoxycorticosterone (3α-5α-

THDOC) in healthy volunteers. Neuropsychopharmacology. 30, 192-195.

13

Esquivel, G., Fernandez-Torre, O, Schruers, K.R.J., Wijnhoven, L.L.W., Griez, E.J.L., 2009. The

effects of opioid receptor blockade on experimental panic provocation with CO2. J.

Psychopharmacol. 23, 975-978.

Gartside, S.E.,, Griffith, N.C., Kaura, V. Ingram, C.D., 2010. The neurosteroid

deidroepiandrosterone (DHEA) and its metabolites altered 5-HT neuronal activity via

modulation of GABAA receptors. Psychopharmacology. 24, 1717-1724

Gorman, J.M., Battista, D., Goetzn, R.R., Dillon, D.J., Liebowitz, M.R., Fyer, A.J., Kahn, J.P.,

Sandberg, D., Klein, D.F., 1989. A comparison of sodium bicarbonate and sodium lactate

infusion in the induction of panic attacks. Arch. Gen. Psychiatry 46, 145-150.

Gorman, J.M., Papp, L.A., Martinez, J., Goetz, R.R., Hollander, E., Liebowitz, M.R., Jordan, F.,

1990. High-dose carbon dioxide challenge test in anxiety disorder patients. Biol. Psychiatry

28, 743-757.

Heydary, B., Le Mellado, J.W., 2002. Low pregnenolone sulphate plasma concentrations in patients

with generalized social phobia. Psychol. Med. 32, 929-933.

Hollander, E., Liebowitz, M.R., Gorman, J.M., Cohen, B., Fyer, A., Klein, D.F., 1989. Cortisol and

sodium lactate-induced panic. Arch. Gen. Psychiatry 46, 135-140.

Levin, A.P., Doran, A.R., Liebowitz, M.R., Fyer, A.J., Gorman, J.M., Klein, D.F., Paul, S.M., 1987.

Pituitary adrenocortical unresponsiveness in lactate-induced panic. Psychiatry Res. 21, 23-

32.

Levine, S., 2000. Influence of psychological variables on the activity of the hypothalamic-pituitary-

adrenal axis. Eur. J. Pharmacol. 405, 149-160.

McEwen, B.S., 1998. Protective and damaging effects of stress mediators. N. Engl. J. Med. 338,

171-179.

Patchev, V.K., Hassan, A.H., Holsboer, D.F., Almeida, O.F., 1994. The neurosteroid

tetrahydroprogesterone counteracts corticotropin-releasing hormone-induced anxiety and

alters the release and gene expression of corticotropin-releasing hormone in the

hypothalamus. Neuroscience 62, 265-271.

Paul, S.M., Purdy, R.H., 1992. Neuroactive steroids. FASEB J. 6, 2311-2322.

Perna, G., Cocchi, S., Bertani, A., Arancio, C., Bellodi, L., 1995. Sensitivity to 35% CO2 in healthy

first degree relatives of patients with panic disorder. Am. J. Psychiatry 152, 623-625.

Perna, G., Biffi, S., Liperi, L., Di Bella, E., Favaro, E., Cucchi, M., Bellodi, L., 2003. Reattività alla

CO2 e polimorfismi della regione regolatoria del gene del trasportatore della serotonina nel

disturbo di panico. Ital. J. Psychopathol. 9, 31.

14

Petrowski, K., Herold, U., Joraschky, P., Wittchen, H.U., Kirschbaum, C., 2010. A striking pattern

of cortisol non-responsiveness to psychosocial stress in patients with panic disorder with

concurrent normal cortisol awakening responses. Psychoneuroendocrinology. 35, 414-421.

Pinna, G., Agis-Balboa, R.C., Pibiri, F., Nelson, M., Guidotti, A., Costa, E., 2008. Neurosteroid

biosynthesis regulates sexual dimorphic fear and aggressive behavior in mice. Neurochem.

Res. 33 (10) 1990-2007

Pisu, M.G., Serra, M., 2004. Neurosteroids and neuroactive drugs in mental disorders. Life Sci. 74,

3181-3197.

Pols, H., Zandberg, J., de Loof, C., Griez, E., 1991. Attenuation of carbon dioxide-induced panic

after clonazepam treatment. Acta Psychiatr. Scand. 84, 585-586.

Purdy, R.H., Morrow, A.L., Plinn, J.R., Paul, S.M., 1991. Synthesis, metabolism and

pharmacological activity of 3α-hydroxysteroids which potentiate GABA-receptor-mediated

chloride ion uptake in rat cerebral cortical synaptoneurosomes. J. Med. Chem. 33, 1572-

1581.

Rupprecht, R., 2003. Induced panic attacks shift γ-aminobutyric acid type A receptor modulatory

neuroactive steroid composition in patients with panic disorder: preliminary results. Arch.

Gen. Psychiatry 60, 161-168.

Rupprecht, R., Holsboer, F., 1999. Neuroactive steroid mechanisms of action and

neuropsychopharmacological perspectives. Trends Neurosci. 22, 410-416.

Rupprecht, R., Di Michele, F., Herman, B., Strohle, A., Lancel, M., Romeo, H., Holsboer, F., 2001.

Neuroactive steroids: molecular mechanisms of action and implications for

neuropsychopharmacology. Brain Res. Rev. 37, 59-67.

Schmidt, N.B., Zvolensky, M.J., 2007. Anxiety sensitivity and CO2 challenge reactivity as unique

and interactive prospective predictors of anxiety pathology. Depression Anxiety 24, 527-

536.

Seier, F.E., Kellner, M., Yassouridis, A., Heese, R., Strian, F., Wiedemann, K., 1997. Autonomic

reactivity and hormonal secretion in lactate-induced panic attacks. Am. J. Physiol. 272,

2630-2638.

Serra, M., Madau, P., Chessa, M.F., Caddeo, M., Sanna E., Trapani, G., Franco,M., Liso, G., Purdy,

R.H. Barbaccia, M.L., Biggio, G., 1999. 2-Phenyl-imidazo[1,2-a]pyridine derivative as

ligands for peripheral benzodiazepine receptors: stimulation of neurosteroid synthesis and

anticonvulsant action in rat. Brit. J. Pharmacol. 127, 177-187

Semeniuk, I., LeMelledo, J.M., Shangri, G.S, 2001.Neuroactive steroid levels in patients with

generalized anxiety disorders. J. Neuropsychiatry Clin. Neurosci. 13, 396-398.

15

Sheehan, D.V., Lecrubier, Y., Janays, J., Knapp, E., Werler, E., Bonora, P., Sheedan, M., Amorim,

P., Baker, M., Sheehan, K.H., Lepine, J.P., 1994. MIINI-International Neuropsychiatric

Interview, Paris-Tampa.

Spivak, B., Maayan, R., Kotler, M., Mester, R., Gil-Ad, R., Shtaif, B., Wizman, A., 2000. Elevated

circulatory level of GABAA-antagonistic neurosteroid{s?} in patients with combat-related

post-traumatic stress disorder. Psychol. Med. 30, 1227-1231.

Strohle, A., Romeo, E., Di Michele, F., Pasini, A., Yassouridis, A., Holsboer, F. Rupprecht, R.,

2002. GABAA receptor-modulating neuroactive steroid composition in patients with panic

disorder before and during paroxetine treatment. Am. J. Psychiatry 159, 145-147.

Strohle, A., Romeo, E., Di Michele, F.A., Pasini, A., Hermann, B., Gajewsky, G., Holsboer, R.,

Rupprecht, R., 2003. Induced panic attacks shift gamma-aminobutyric acid type A receptor

modulatory neuroactive steroid composition in patients with panic disorder. Preliminary

results. Arch. Gen. Psychiatry 60, 161-168.

Tait, G.R., McManus, K., Bellavance, F., Lara, N., Chrapko, W., Le Melledo, J.M., 2002.

Neuroactive steroid changes in response to challenge with the panicogenic agent

pentagastrin. Psychoneuroendocrinology 27, 417-429.

Toru, I., Aluoja, A., Vohma, U., Raag, M., Vasar, V. Maron, E., Shlik, J., 2010. Association

between personality traits and CCK-4-induced panic attacks in healthy volunteers.

Psychiatry Res. 178, 342-347.

Ursin, H., Baade, E., Levine, S., 1978. Psychobiology of Stress. Academic Press, New York.

Van Duinen, M.A., Schruers, K.K., Jaegers, E., Maes, M., Griez, E.J., 2004. Salivary cortisol in

panic: Are males more vulnerable? Neuroendocrinol. Lett. 25, 386-390.

Wieland, S., Lan, N.C., Miradeseghi, S., Gee, K.W., 1991. Anxiolytic activity of the progesterone

metabolite 5α-pregnan-3α-ol-20-one. Brain Res. 565, 263-268.

Woods, S.W., Charney, D.S., McPerson, C.A., Gradman, A.H., Heninger, G.R., 1987. Situational

panic attacks: behavioural, physiologic and biochemical characterization. Arch. Gen.

Psychiatry 44, 365-375.

Zinbarg, R.E., Brown, T.A., Barlow, D.H., Rapee, R.M., 2001. Anxiety sensitivity, panic, and

depressed mood: a reanalysis teasing apart the contribution of the two levels in the

hierarchical structure of the Anxiety Sensitivity Index. J. Abnorm. Psychol. 10, 172-177.

Zwanzger, P., Baghai, T.C., Schuele, C., Strohle, A., Padberg, F., Kathmann, N., Schwarz, M.,

Moller, H.J., Rupprecht, R., 2001. Vibagastrin decreases cholecystokinin-tetrapeptide

(CCK-4) induced panic in healthy volunteers. Neuropsychopharmacology. 25, 699-703.

16

Zwanzger, P., Eser, D., Padberg, F., Baghai, T.C., Schule, C., Rupprecht R., 2003. Neuroactive

steroids are not affected by panic induction with 50 μg cholecystokinin-tetrapeptide (CCK-

4) in healthy volunteers. J. Psychiatr. Res. 38, 215-217

17

Figure Legends

Figure 1: Effect of CO2 inhalation on the plasma concentrations of progesterone (A), 3α,5α-

tetrahydroprogesterone (B), 3α,5α-tetrahydrodesoxycorticosterone (C), dehydroepiandrosterone (D)

and cortisol (E) in the responders, 15 minutes before (T-15), immediately before (T0) and 20 minutes

later, immediately after (T20) CO2 administration, then 10 min (T30), 30 min (T50), and 50 min (T70)

after the end of CO2 inhalation.

Data are means ± SEM.

Figure 2.: Effect of CO2 inhalation on the plasma concentrations of progesterone (A), ), 3α,5α-

tetrahydroprogesterone (B), 3α,5α-tetrahydrodesoxycorticosterone (C), dehydroepiandrosterone (D)

and cortisol (E) in non-responders, 15 minutes before (T-15), immediately before (T0) and 20

minutes later, immediately after (T20) CO2 administration, then 10 min (T30), 30 min (T50), and 50

min (T70) after the end of CO2

18

TABLE 1.

GLM ANALYSIS FOR REPEATED MEASURES ON NEUROSTEROID RESPONSES TO CO2

INHALATION

Between Subjects Effects Post-hoc

Progesterone Group F(1.3)=18.354; p<0.001 1>2.3.4

  Response F(1.3)=0.415; p=0.522  

Allopregnanolone Group F(1.3)=1.576; p=0.206  

  Response F(1.3)=0.660; p=0.420  

3α,5α-THDOC Group

F(1.3)=0.1.902;

p=0.143  

  Response F(1.3)=0.222; p=0.640  

dehydroepiandrosterone Group F(1.3)=2.047; p=0.058  

  Response F(1.3)=0.471; p=0.496  

cortisol Group F(1.3)=0.768; p=0.517  

  Response F(1.3)=1.007; p=0.320  

Group: women in the follicular phase, women in the luteal phase, men, women on

contraceptive pills

Response = responders and non responders to CO2 inhalation

19

TABLE 2.

VAS-A AND PSL-III-R SCORES BEFORE AND AFTER CO2 INHALATION

PAIRED SAMPLE T-TEST.

t0 t1    

   Mean

Std.

Deviation Mean

Std.

Deviationt P

VAS-A VASPRE 6.65 9.4 44.41 31.1 -8.578 <0.001

PSL-III-

R

Breath .78 1.1 1.95 1.2 -4.750 <0.001

Dizziness .44 .8 1.14 1.2 -3.050 .004

Palpitation .29 .6 2.12 1.3 -9.104 <0.001

Trembling .00 .00 .90 1.1 -5.127 <0.001

Sweating .07 .3 1.43 1.1 -7.993 <0.001

Choking .10 .3 1.98 1.4 -9.173 <0.001

Nausea .07 .3 .61 .9 -3.516 0.001

Unreality .00 .0 .19 .5 -2.715 0.010

Numbness .14 .5 .43 .7 -3.106 0.003

Hot flush .12 .3 1.55 1.1 -8.908 <0.001

Chest discomfort .00 .0 .62 .9 -4.698 <0.001

Dying .02 .2 .27 .6 -2.504 0.016

  Control loss .00 .0 .27 .6 -2.899 0.006

t0: before CO2 inhalation; t1: after CO2 inhalation

20

FIGURE 1.

GLM REPEATED MEASURES FOR PROGESTERONE, ALLOPREGNANOLONE, 3α,5α-

TETRAHYDRODESOXICORTICOSTERONE, DEHYDROEPIANDROSTERONE AND

CORTISOL AMONG RESPONDERS

21

22

Effect of CO2 inhalation on the plasma concentrations of Progesterone, Allopregnanolone, 3α,5α-

THDOC, Dehydroepiandrosterone and Cortisol for each group across the time. Repeated measures

of each hormone: 15 minutes before CO2 inhalation (t-15), immediately before (t0) and 20 (t20),

30 (t30), 50 minutes (t50) and 70 (t70) minutes after the beginning of CO2 inhalation.

23

FIGURE 2.

GLM REPEATED MEASURES FOR PROGESTERONE, ALLOPREGNANOLONE, 3α,5α-

TETRAHYDRODESOXICORTICOSTERONE, DEHYDROEPIANDROSTERONE AND

CORTISOL AMONG NON RESPONDERS

24

25

Effect of CO2 inhalation on the plasma concentrations of Progesterone, Allopregnanolone, 3α,5α-

THDOC, Dehydroepiandrosterone and Cortisol for each group across the time. Repeated measures

of each hormone: 15 minutes before CO2 inhalation (t-15), immediately before (t0) and 20 (t20),

30 (t30), 50 minutes (t50) and 70 (t70) minutes after the beginning

26