All-trans retinoic acid-induced hypothalamus-pituitary-adrenal hyperactivity involves glucocorticoid receptor dysregulation

In addition to activating RARs, RA also serves as ligand for the nuclear receptor peroxisome proliferator-activated receptor β/δ (PPAR β/δ).^1 As distinct activities of RA may be mediated through alternative signaling mechanisms of RAR and PPAR β/δ,^2-4 activation of each receptor might produce divergent effects on the HPA axis. Therefore, we compared the effects on plasma CORT concentration changes during acute infusion of RAR-α-selective agonist and antagonist, and PPAR β/δ-selective agonist as well.

For the acute drug administration experiment, 30 rats were randomly assigned to 5 groups (n=6 each group): vehicle (VEH), RA, RAR-α-selective agonist (AM580), PPAR β/δ-selective agonist (GW0742), RA plus RAR-α specific antagonist (RA+ER50891).

Drug administration and blood sampling

In the acute drug treatment experiment, animals were connected to sampling lines attached to a metal collar (attached 12 h before experiment for adaptation) with a counter-balanced beam. Injection speed was controlled by a pump. Drugs were given between 08:00-10:30 am. 20 μg RA (Sigma-Aldrich, St. Louis, MO) was dissolved in 20 μl 1:1 v/v DMSO: saline while avoiding light exposure. The VEH group received 1:1 v/v DMSO: saline. AM580 (Tocris, Bristol, UK) was infused at a 30 nM dose.^7-9 For GW0742 (Tocris, Bristol), the infusion dose was 100 nM. To ensure complete blockade, ER50891 (Tocris, Bristol) was infused at a dose 100 times that of RA (0.333 M). All drugs were dissolved in 1:1 v/v DMSO: saline and infused at a volume of 20 μl.

After a basal blood sample was collected at t=0 min, the animals received an initial bolus infusion of the drug for 5 min at 1 μl/min. Then infusion speed was then re-adjusted to 7.5 μl/h for the following 2 h. Subsequent samples were sequentially collected at t=15 min, 30 min, 60 min and 120 min.

Open-field tests

Open-field tests analyze spontaneous exploratory activity and curiosity to novel environment. The open field test and the elevated plus maze (EPM) are often combined together to measure anxiety-related behaviors. The apparatus consists of a black arena of 100 cm × 100 cm, with a 30 cm high black wall. The floor marked with white grids (6 mm) was divided into 16 equal-size squares. Rat was placed in the center (central 4 squares) for 5 min. The duration and latency after which each rat left the center and the distance it traveled in the center were analyzed by video camera and processed using EthoVision (Noldus, Wageningen, The Netherlands).

To characterize the specificity of antibodies used in immunohistochemistry, western blot was performed.^15

Details on specificity of the primary antibody for CRH used in the current study are given in our previous studies.^5 We further performed Western blot in rat hypothalamus tissue that was probed with the anti-RAR-α (rabbit; Abcam, UK) and anti-c-fos (rabbit; Santa Cruz, CA) primary antibody recognizing protein bands around 52 kDa and 62 kDa respectively. As a negative control, another blot was incubated without primary antibody. Protein samples were detected with HRP-conjugated goat-anti-rabbit IgG (1:2000; Promega, Madison, WI) followed by detection with ECL chemiluminescence (SuperSignal West Pico chemiluminescent Substrate, Pierce, IL). In each trial of our immunohistochemical experiments, a negative control sample was included.

Quantification of immunohistochemical data

Immunohistochemical staining of each marker was performed in every 8th section taken along the rostrocaudal axis throughout PVN. We collected 40 μm thick sections of the PVN region starting from bregma -0.92 mm until bregma -2.12 mm which yielded 4 sections for each marker, that were used for quantification.

The numbers of CRH, c-fos, RAR-α, and GR immunopositive cells in the PVN were counted manually and the main principle and procedures have been described previously.^16 Briefly, the PVN was outlined manually at a low magnification (10× objective) before a total of 35 gridfields were superimposed on this image and cell numbers assessed (shown in Fig S9). At higher magnification (40× objective), cell numbers were confirmed in a selection of retrieved fields. The number from each field was summed up together before multiplied by 8 to estimate the total immunopositive cell number in the PVN. All cell counting was done by investigator blind to the experimental condition.

Statistics

Data were expressed as mean ± SEM. For evaluation of CORT values in acute experiment, a repeated measure ANOVA was used. One-way ANOVA was used to compare time effects within each group. For evaluation of body weight change, data were subjected to a one-way ANOVA followed by post-hoc Tukey test. For evaluation of basal CORT values, behavioral and immunohistochemistry results, data were also evaluated by a two-factor ANOVA to study the main effects of RA and MIF treatment and possible interaction between the two factors.

As shown in Fig 1B, under basal conditions, significantly different CORT levels were found by RA (F(1,16)=6.902, p=0.018) but not MIF (F(1,16)=2.442, p=0.138) treatment . A significant interaction between RA×MIF (F(1,16)=11.405, p=0.004) was found.

Difference in sucrose uptake was found among the four groups by RA (F(1,18)=4.452, p=0.049) but not MIF (F(1,18)= 1.085, p=0.311) treatment. A significant interaction between RA×MIF (F(1,18)=9.529, p=0.006) was found (Fig 1D).

In the EPM experiment, no significant difference was found in entry frequency in open arms among the four groups either by RA (F(1,19)=1.225, p=0.282) or MIF (F(1,19)=3.774, p=0.067) treatment. No interaction between RA×MIF (F(1,19)=4.055, p=0.058) was found (Fig 1E). For the duration percentage in open arms, no significant difference was found among the four groups either by RA (F(1,20)=2.414, p=0.136) or MIF (F(1,20)=1.921, p=0.181) treatment. A significant interaction between RA×MIF (F(1,20)=4.055, p=0.022) was found (Fig 1F).

As shown in Fig 3E, for CRH-IR cell number, significant interaction between RA×MIF (F(1,20)=8.568 p=0.008) was found among the four groups. However, no significant difference was found either by RA (F(1,20)=2.407, p=0.136) or MIF (F(1,20)=1.936, p=0.179) treatment (Fig 3E). A significantly different c-fos-IR cell number was found among groups by RA (F(1,20)=5.593, p=0.028) but not MIF (F(1,20)=3.473, p=0.077) treatment. Significant interaction between RA×MIF (F(1,20)=4.593, p=0.045) was found (Fig 3F). For RAR-α-IR cell number, significant difference was found among groups by both RA (F(1,20)=28.900, p<0.001) and MIF (F(1,20)=20.962, p<0.001) treatment. No interaction between RA×MIF (F(1,20)=4.179, p=0.054) was found (Fig 3G). Also, significantly different GR-IR cell number was found among groups by RA (F(1,20)=4.966, p=0.037) but not MIF (F(1,20)=0.999, p=0.330) treatment. A significant interaction between RA×MIF (F(1,20)=9.740, p=0.005) was found (Fig 3H).

We first compared effects of acute drug infusion on plasma CORT in the following 5 groups: VEH, RA, RAR-α selective agonist AM580, PPAR β/δ selective agonist GW0742 and RA plus RAR-α selective antagonist ER50891 (Fig S1).

Plasma CORT concentrations were significantly different among groups (time effect, F(4,84)=25.527, p<0.001; group effect, F(4,21)=43.645, p<0.001; time×group effect (F(16,84)=11.690, p<0.001). Both RA (p<0.001) and AM580 (p<0.001) elevated CORT concentrations significantly compared to VEH, whereas no significant difference was found between GW0742 (p=0.999) and VEH groups. The RAR-α selective antagonist ER50891 completely blocked the RA effect (p=0.977 RA+ER50891 vs VEH group).

After RA infusion, at 60 min (p=0.008) and 120 min (p<0.001), CORT levels were significantly elevated compared to t=0 min. Interestingly, CORT concentration increased earlier after AM580 infusion, with a sharp increase at 15 min (p=0.034) and continued to increase at 30 min (p=0.003), 60 min (p=0.001) and 120 min (p<0.001).

No significant different body weight percentual change was found among the following four chronic treatment groups (Fig S2) (F(3,19)=0.189, p=0.902): vehicle (VEH) group, all-trans retinoic acid (RA) group, vehicle plus mifepristone (VEH+MIF) group and all-trans retinoic acid plus mifepristone (RA+MIF) group.

In the open field experiment, no significant difference in inner distance (Fig S3A, F(3,17)=1.694, p=0.206), duration (Fig S3B, F(3,20)=0.384, p=0.765) or latency (Fig S3C, F(3,20)=0.082, p=0.969) was found among the four chronic treatment groups: vehicle (VEH) group, all-trans retinoic acid (RA) group, vehicle plus mifepristone (VEH+MIF) group and all-trans retinoic acid plus mifepristone (RA+MIF) group.

No significant difference in inner distance was found among the four groups by either RA (F(1,17)=0.413, p=0.529) or MIF (F(1,17)=3.922, p=0.064) treatment. No interaction between RA×MIF (F(1,17)=1.578, p=0.226) was found (Fig S3A). Again, no significant difference in inner duration was found among groups either by RA (F(1,19)=0.102, p=0.753) or MIF (F(1,19)=1.259, p=0.276) treatment. No interaction between RA×MIF (F(1,19)=0.074, p=0.789) was found (Fig S3B). Also, no significant difference in inner latency was found among the four groups either by RA (F(1,20)=0.140, p=0.713) or MIF (F(1,20)=0.054, p=0.819) treatment. No interaction between RA×MIF (F(1,20)=0.057, p=0.814) was found (Fig S3C).
Supplementary Figure 4

Immunostaining patterns of CRH staining in the hypothalamic paraventricular nucleus (PVN) compared among different chronic treatment groups were shown in Fig S4: (A) vehicle (VEH) group, (B) all-trans retinoic acid (RA) group, (C) vehicle plus mifepristone (VEH+MIF) group and (D) all-trans retinoic acid plus mifepristone (RA+MIF) group.

Immunostaining patterns of c-fos staining in the hypothalamic paraventricular nucleus (PVN) compared among the four chronic treatment groups were shown in Fig S5: (A) vehicle (VEH) group, (B) all-trans retinoic acid (RA) group, (C) vehicle plus mifepristone (VEH+MIF) group and (D) all-trans retinoic acid plus mifepristone (RA+MIF) group.

Immunostaining patterns of RAR-α staining in the hypothalamic paraventricular nucleus (PVN) compared among the four chronic treatment groups were shown in Fig S6: (A) vehicle (VEH) group, (B) all-trans retinoic acid (RA) group, (C) vehicle plus mifepristone (VEH+MIF) group and (D) all-trans retinoic acid plus mifepristone (RA+MIF) group.

Immunostaining patterns of GR staining in the hypothalamic paraventricular nucleus (PVN) compared among different chronic treatment groups were shown in Fig S7: (A) vehicle (VEH) group, (B) all-trans retinoic acid (RA) group, (C) vehicle plus mifepristone (VEH+MIF) group and (D) all-trans retinoic acid plus mifepristone (RA+MIF) group.

Western blots probed with anti-RAR-α and anti-c-fos primary antibody each revealed a single band of the expected molecular weight approximately 52 kDa and 62 kDa, while omission of the primary anti-RAR-α and anti-c-fos antibody (negative control, NC) failed to show any bands (Fig S8). Due to the relatively small molecular weight of CRH (around 21 kDa), we refer for specificity validation to our previous study.^5 The same applies for the anti-GR antibody that has been validated before.^17

Together, we can confirm the specificity of the antibodies used in our immunohistochemical study.
Supplementary Figure 9

As shown in Fig S9, the whole PVN was outlined manually at low magnification (100× magnification, 10 x objective) before 35 grids of fields were superimposed. Cell number from each selected grid field was counted.

(P.Hu et al., “All-trans retinoic acid-induced hypothalamus-pituitary-adrenal hyperactivity involves glucocorticoid receptor dysregulation”)

Plasma CORT concentration changes at different time-points during a 2-hour intracerebroventricular (i.c.v.) infusion were compared among vehicle (VEH), all-trans retinoic acid (RA), RAR-α selective agonist AM580, PPAR β/δ selective agonist GW0742, and all-trans retinoic acid plus RAR-α specific antagonist ER50891 (RA+ ER50891) groups.

Data are presented as mean ± SEM (n=6 animals per group). * p<0.05; ** p<0.01: significant different compared with t=0 min within group.
Supplementary Figure 2: Comparison of body weight percentual change among the four chronic intracerebroventricular (i.c.v.) treatment groups: vehicle (VEH) group, all-trans retinoic acid (RA) group, vehicle plus mifepristone (VEH+MIF) group and all-trans retinoic acid plus mifepristone (RA+MIF) group.

Data were presented as mean ± SEM (n=6 animals per group).

Data were presented as mean ± SEM (n=6 animals per group).

CRH immunostaining patterns of (A) vehicle (VEH), (B) all-trans retinoic acid (RA), (C) vehicle plus mifepristone (VEH + MIF), and (D) all-trans retinoic acid plus mifepristone (RA + MIF) groups were shown respectively. 200× magnification. Black dotted line highlighted the border region of single side of PVN. Scale bar: 50 μm. Arrowhead points to the example CRH-immunopositive cell shown in the center of higher magnification inset as in Fig 3A.

c-fos immunostaining patterns of (A) vehicle (VEH), (B) all-trans retinoic acid (RA), (C) vehicle plus mifepristone (VEH + MIF), and (D) all-trans retinoic acid plus mifepristone (RA + MIF) groups were shown respectively. 200× magnification. Black dotted line highlighted the border region of single side of PVN. Scale bar: 50 μm. Arrow points to the example c-fos-immunopositive cell region shown in the bottom-left corner of higher magnification inset as in Fig 3B.

RAR-α immunostaining patterns of (A) vehicle (VEH), (B) all-trans retinoic acid (RA), (C) vehicle plus mifepristone (VEH + MIF), and (D) all-trans retinoic acid plus mifepristone (RA + MIF) groups were shown respectively. 200× magnification. Black dotted line highlighted the border region of single side of PVN. Scale bar: 50 μm. Arrow points to the example RAR-α-immunopositive cell shown in the upper-left corner of higher magnification inset as in Fig 3C.

GR immunostaining patterns of (A) vehicle (VEH), (B) all-trans retinoic acid (RA), (C) vehicle plus mifepristone (VEH + MIF), and (D) all-trans retinoic acid plus mifepristone (RA + MIF) groups were shown respectively. 200× magnification. Black dotted line highlighted the border region of single side of PVN. Scale bar: 50 μm. Arrow points to the example GR-immunopositive cell shown in the bottom-left corner of higher magnification inset as in Fig 3D.

Western blot example of RAR-α (A) and c-fos (B) protein bands in the hypothalamus of chronic vehicle (VEH)-treated rats. Arrows point to the molecular weight bands of markers showing relative position of RAR-α (52 kDa), and c-fos (62 kDa) respectively. No stainings were detected when incubation with primary antibody was omitted (negative control, NC).

An example of PVN with immunostaining of RAR-α was shown. PVN was outlined manually at low magnification (100× magnification, 10× objective). 35 grids of fields were superimposed. Cell number from each selected grid field was counted. Black dotted line highlighted the border region of the PVN.