Mimecan Regulates Corticosterone Secretion and Plays A Critical Role in Adrenal Responses to Stress

Background: A functional hypothalamic-pituitary-adrenal (HPA) axis is critical for host defenses to outside stimuli. The adrenal cortex is seemingly endowed with distinct functional units that are regulated by adrenocorticotropic hormone (ACTH). We have found that mimecan, a small leucine-rich proteoglycan expressed in the adrenal gland, has yet to be characterized in functional terms. Results: Herein, we have demonstrated the following properties: 1) adrenal mimecan expression in mouse models is signicantly downregulated under hypoglycemia and scalded stress; 2) expression of mimecan in adrenal cells may be downregulated through ACTH or upregulated by glucocorticoid via related receptors (GRs); and 3) mimecan stimulates corticosterone secretion in adrenal tissues. The latter was proven using in vivo and in vitro studies to conrm the ACTH-independent activity of mimecan-maltose-binding protein (-MBP). Relative to litter-mate mice, the basal-state diurnal rhythm of corticosterone secretion is disrupted in mimecan knockout mice, and corticosterone secretion is increased under restraint stress conditions. Conclusions: These ndings offer the rst evidence that mimecan is key in regulating the HPA axis, assuming a critical role in adrenal responses to stress. ANX FF, GE Healthcare) chromatography. Purity of the mimecan-MBP fusion protein was 96%, as determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and silver staining. MBP was expressed and puried (98%) for use as a control.

secretion is disrupted in mimecan knockout mice, and corticosterone secretion is increased under restraint stress conditions. Conclusions: These ndings offer the rst evidence that mimecan is key in regulating the HPA axis, assuming a critical role in adrenal responses to stress.

Background
Each organism must maintain a complex dynamic equilibrium, otherwise known as homeostasis. Stress is a state in which homeostasis is threatened by emotional or physical stressors, although various physiologic and behavioral adaptive responses may be restorative in this regard (1). Exposure to stressful challenges incites behavioral and physical changes that are normally adaptive and limited over time, improving chances for survival. These responses must be appropriate in magnitude and duration; otherwise, they may have detrimental effects on numerous physiologic functions of organism, leading to a state of disease-causing disturbed homeostasis.
The HPA axis is pivotal in responses to stress. The HPA axis activation is initiated by the activation of parvocellular neurons in the paraventricular nucleus (PVN) of the hypothalamus and the release of corticotropin-releasing hormone (CRH) and arginine vasopressin (AVP) from the external zone of the median eminence into portal circulation (2,3). CRH acts upon anterior pituitary, stimulating synthesis and secretion of adrenocorticotropic hormone (ACTH), which then stimulates adrenal cortex to secrete glucocorticoid (GC) (cortisol in humans; corticosterone in rodents) (2,3). A primary effect of stressinduced glucocorticoid release is the inhibition of ongoing HPA-axis activation through negative feedback at the level of hypothalamus and pituitary and at upstream limbic structures, thus inhibiting ACTH and CRH secretion through mineralocorticoid (MR) and glucocorticoid receptors (GR) (2,3).
HPA-axis dysfunction is implicated in the pathogenesis of various stress-related physical and psychological diseases, such as Cushing' syndrome, panic disorder, and post-traumatic stress disorder, all of which re ect heightened HPA-axis activity. On the other hand, adrenal insu ciency, chronic fatigue syndrome, and atypical depression are associated with reduced activity of the HPA axis (4,5). Hence, mechanisms regulating functions of the HPA axis are of utmost importance in managing stress-related disease. Such mechanisms unfortunately have yet to be fully clari ed.
Originally isolated from bone, mimecan (osteoglycin) belongs to the family of small leucine-rich proteoglycans (SLRPs) (6). SLRPs are abundant in bone matrix, cartilage cells, and connective tissues. They are also essential for collagen brillogenesis and are central to cellular growth, differentiation, and migration (7). Although the mimecan gene encodes a 34-kDa full-length protein, a 12-kDa mature protein corresponding to the 105 carboxyl-terminal amino acids of mimecan has been isolated from bovine bone. A 25-kDa keratan sulfate glycoprotein corresponding to the 223 carboxyl-terminal amino acids of mimecan has been isolated from bovine cornea (8).
To date, the physiologic functions of mimecan remain obscure. We have previously cloned the full-length cDNA of human mimecan (accession number: AF100758) (9), establishing mimecan as a novel satiety hormone in adipose tissue propagated by IL-1β and IL-6 induction within hypothalamus (10). Our previous studies have also shown that mimecan is expressed in human pituitary corticotroph cells and the AtT-20 mouse corticotroph cell line and mimecan gene expression in pituitary corticotroph cells is up-regulated by glucocorticoid (GC) in a time-and dose-dependent manner (9,11). Besides, mimecan stimulated adrenocorticotrophic hormone (ACTH) secretion in pituitary corticotroph AtT-20 cells (12). Because mimecan is important in the functions of HPA axis, the present study was undertaken to further explore the expression of mimecan in adrenal glands and its effects on HPA axis functions.

Animals
Wild-type C57BL/6 mice were bought from Shanghai Experimental Animal Center, Chinese Academy of Sciences. Mimecan knockout mice (mim -/-) were generated as described elsewhere (11,12). The animals were housed in a temperature-controlled room (23º C) subject to12-h light/dark cycles, allowing ad libitum access to chow and water. Adult male (8-10 week-old) mim -/mice and age-matched wild-type C57BL/6 mice were used for experiments. All animal-related investigations were conducted in accordance with our institutional guidelines on ethical animal care and were approved by the Animal Care and Use Committee in the Ninth Hospital A liated to Shanghai Jiaotong University School of Medicine. Hypoglycemic-mouse model C57BL/6 mice were randomly assigned to stressed or control group of each time point 0h, 1h, 2h, 4h, 6h, 8h, and 12h (10 mice for each time point). C57BL/6 mice of stressed group were fasted overnight for 12 h. Blood glucose levels of 0h were assayed in blood obtained by tail cuts at 8 am using a glucometer (Accu-Chek Compact; Roche Diagnostics), and intraperitoneal (IP) injections of insulin (3 IU/kg) were delivered at 9 am(13). The mice were subsequently sacri ced by cervical dislocation after blood glucose measure in tail vein at 1h, 2h, 4h, 6h, 8h and 12h after injection and adrenal glands and lung tissues were immediately taken and frozen at -80 ºC. The target blood glucose concentration was <40 mg/dl, a value conventionally used to de ne hypoglycemia during development. The other half of the litter was injected with the equivalent volume of 0.9% saline (control group). Scalded-mouse model C57BL/6 mice were randomly assigned to scalded or control group (15 mice per group) 24 hours after the hairs of mice were removed by using hair removal agents made of barium sul de, mice were randomly divided into stressed group and control group. Both groups of mice were anesthetized with 2.5% pentobarbital sodium at the dose of 35mg/kg and placed on boards designed for 10% of total body surface area exposure and immersed (8 seconds) in a water bath held at 90°C (stress group) or room temperature (control group) (14). This method delivers a full-thickness cutaneous burn as con rmed by histological examination. Four hours after burn or sham injury, the mice were fully recovered from anesthesia and were sacri ced by cervical dislocation. Portions of scalded skin and unilateral were xed in buffered formalin for morphologic evaluation. Part samples of adrenal glands and lung tissues were formalin-xed and para n-embedded and part were immediately frozen and stored at -80°C until analyzed.
Restraint stress model 5 mim -/mice and 5 wild-type litter mates were restrained using plastic tubes with a narrow end that exposed the mouse's head, obtaining blood (by tail cuts) after 60 min of restraint and being released for 60 min.
ACTH stimulation test C57BL/6 mice were randomly assigned to each group (12 mice for each time point). We administered ACTH (0.085 IU/g body weight) in mice by IP injection (9:00 am), then sacri cing the mice by cervical dislocation at the indicated time points 0h, 2h, 4h, 6h, 8h, and 12h. Adrenal glands were immediately taken and frozen at -80. ACTH (sigma) were diluted by sterile 0.9% saline and placed on ice to remain effective. Corresponding control groups were stimulated using equivalent volume of 0.9% saline. In mim -/mice and wild-type litter mates (21 mim -/mice and 15 WT mice), same dose of ACTH was given by IP injection at 9am. Blood samples for corticosterone measurement were removed from the tail vein before and 15min, 30min, 45min, 60min, and 120min after ACTH administration.
Dexamethasone suppression test C57BL/6 mice were randomly assigned to each group (10 mice per group for each time point). Dexamethasone (DEX, sigma, 0.05 ug/g body weight) or 0.9% saline were administered by intramuscular (IM) injection, mice were sacri ced by cervical dislocation at the indicated time points 2h, 12h and 36h at 7:30pm (when cortisol secretion peaks) and adrenal glands were immediately taken and frozen at -80. In mim -/mice and wild-type litter mates (21 mim -/and 15 WT ), same dose of DEX was given by IM injection. Blood samples for corticosterone and ACTH measurement were removed from the tail vein 5h, 8h, 24h and 32h after DEX administration at 12am.
Circadian rhythm determination of corticosterone 11 mim -/mice and 12 wild-type litter mates were used, and blood was taken from tail vein every 4 h for serial serum corticosterone determinations.
Mimecan-MBP stimulation C57BL/6 mice were administered with 0.1 μmol/kg mimecan-MBP, 0.1 μmol/kg MBP, or an equivalent volume of 0.9% saline by IP injection at 9am. Mice were sacri ced by decapitation and trunk blood was obtained 0.5h, 2h, 24h and 48h after injection for hormonal analysis. For 24h and 48h group, injection was repeated every 8 hours.
Primary adrenal gland cell isolation and cell culture Primary adrenal gland cells were isolated by collagenase digestion method. Adrenal glands were obtained from 20 adult male C57BL/6 mice after decapitation. Tissues were washed with HBSS. Pieced sliced fragments were dispersed in preparation buffer containing 30mg/ml type 1 collagenase (Sigma).
Dispersed cells were centrifuged and resuspended in F-12K medium containing 15% horse serum and 2.5% fetal calf serum. Then, cells were distributed in 12-well plates and incubated at 37°C under 5% CO2 for 12 h until they were used.
The Y-1 mouse adrenocortical tumor cell line was obtained from American Tissue Type Collection (ATCC, VA, USA), which is a subclone of the corticotropin-responsive cell line originally developed by Yasumura et al. (15). The cell line was maintained in F-12K medium containing 15% horse serum and 2.5% fetal calf serum (FCS) (GIB-CO, USA) in a 5% CO2-humidi ed atmosphere at 37°C. All cell cultures were routinely passaged at 90-95% con uency. Before the experiment, cells were preincubated with F-12K medium containing 0.2% BSA for24 h. Then treated with ACTH (sigma), DEX (sigma), MBP-mimecan fusion protein made in our lab (10), speci c cell pathway inhibitors, and so on. Medium samples were collected and stored at-80•C at the end of the experiments for hormone content analysis. Cells in the culture plates were processed for RNA extraction as indicated below.
Quantitative real-time polymerase chain reaction (PCR) Gene expression was assessed by relative quanti cation (2 ΔΔCt method), using an ABI Prism 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA), 96-well plates, and SYBR Premix Ex Taq (Takara Bio, Shiga, Japan) according to the manufacturer's instructions (16). All samples were normalized to values of β-actin; results were expressed as fold-changes of threshold cycle (CT) values relative to controls. Cycling parameters were 95ºC for 10 s, then 40 cycles of 95ºC for 5 s and 60ºC for 31 s. Analysis was done in quadruplicate, repeating experiments independently three times. Primers were showed in Supplementary Table 1.
Blood collection and hormone assays Whole blood was collected into iced empty or heparinized tubes. Blood was centrifuged at 2000 g for 20 min at 4 ºC, and then plasma or serum were recentrifuged at 6000 g for 10 min at 4 ºC and stored at -80 ºC for subsequent determination of ACTH in plasma and corticosterone in serum. ACTH or corticosterone concentrations of the mouse plasma or serum and Y-1 cell culture media were measured by ELISA commercially available kits (EK-001-21, Phoenix pharmaceuticals, USA and Cayman, USA) as described by the manufacturer.
Fusion protein puri cation and antibody production Mimecan-MBP fusion protein and MBP protein puri cation was conducted as described earlier by our group (10). The cDNA encoding 12 kDa human mimecan (residues 175-279) was subcloned into pGEX-5X-2 (GE Healthcare) and overexpressed in Escherichia coli BL21 (DE3) cells. Puri ed mimecan-MBP fusion protein was used for antibody production. Rabbits and mice were immunized with recombinant protein in Freund's adjuvant (Sigma-Aldrich, St. Louis, MO, USA) for polyclonal and monoclonal antibody production, respectively. Antibodies were puri ed using Protein G (GE Healthcare). The monoclonal subtype was identi ed as IgG1-κ. Human cDNA encoding 12 kDa mimecan (residues 175-279) was subcloned into pMAL-c2x (NEB) and overexpressed in BL21 (DE3) cells. Cells were grown at 37 °C to an optical density at 595nm(A595) of 0.6-0.8, induced with 0.5 mM isopropyl-β-D-thiogalactoside (IPTG) for 5 h, and centrifuged. Cells were sonicated, centrifuged, and the fusion protein in the supernatant was puri ed by a nity chromatography (MBPTrap HP, GE Healthcare), gel ltration (Superdex 200, 10/300 GL, GE Healthcare), and ion exchange (HiTrap ANX FF, GE Healthcare) chromatography. Purity of the mimecan-MBP fusion protein was 96%, as determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and silver staining. MBP was expressed and puri ed (98%) for use as a control.

Northern blot
Northern blot was performed using the non-isotopic digoxigenin (DIG) Northern Starter Kit (Roche Diagnostics, Rotkreuz, Switzerland) as directed by the manufacturer (17). Target fragments, mouse mimecan, and mouse steroidogenic acute regulatory protein (StAR) were cloned into PGEM-T Easy vectors and con rmed by restriction enzyme digestion and sequence analysis. DIG-labeled probes were generated by transcription, using SP6/T7 RNA polymerase from the DIG RNA Labeling Kit. Total RNA was isolated from mouse tissues by TRIzol reagent (Invitrogen), and spectrophotometry was used to gauge total RNA content. 10 ug mRNA composed of equal amount of mRNA from 10 mice in a group per lane were applied to a 1.2% agarose-formaldehyde denaturing gel and transferred by capillary blotting to positively charged nylon membranes (Roche). The membranes were then baked at 80°C for 2 h. Hybridization was performed at 68°C, with overnight agitation. The membranes were washed twice for 5 min (room temperature), using 2× standard saline citrate (SSC) and 0.1% SDS, then twice for 15 min (68°C) using 0.1× SSC and 0.1% SDS. Finally, the membranes were washed, blocked, and incubated with anti-DIG serum/alkaline phosphatase conjugate. CDP-Star (Roche) served as the chemiluminescence substrate. Signals were visualized on x-ray lm.

In situ hybridization
Target fragments (Mimecan) were cloned into PGEM-T easy vector (Promega) and con rmed by automated sequencing. The RNA probes were labeled by using the DIG or Fluorescein RNA labeling kit (SP6/T7; Roche). The adrenal glands from the C57BL/6 mouse were cut into serial frozen sections (5 um). These sections were rst xed in 4% paraformaldehyde and digested in 1 ug/ml protein kinase buffer. After prehybridization, the sections were incubated with hybridization solution containing 0. After blocking for 1 h, sections were incubated overnight at 4℃ with Anti-Fluorescein AP-conjugated secondary antibody, after being washed in MABT, they were incubated with Fast red staining buffer in a humidi ed chamber. To terminate the reaction, samples were rinsed several times with nuclease-free water and were visualized by light microscopy or uorescence microscopy Leica .

Statistical analysis
All data were individually expressed as mean ± SD. When statistical analyses were performed, data were compared by one-way ANOVA or Student's t-test, setting statistical signi cance at p<0.05.

Results
Mimecan are mainly expressed in adrenal cortex and medullary mesenchyme.
In this study, we rst examined the distribution of mimecan in adrenal tissues of the C57BL/6 mouse using immunohistochemistry and in situ hybridization. As indicated by the results of polyclonal antimimecan immunostaining, mimecan expressed mainly in adrenal cortex (Fig. 1I-L). This is consistent with results of in situ hybridization using uorescein-labeled mimecan antisense RNA probes (Fig. 1A-B). Adrenal medullary cells are chie y composed of adrenaline-and noradrenaline-releasing chroma n cells, both expressing catecholamine synthesizing enzymes, such as tyrosine hydroxylase (TH) and dopamine β-hydroxylase (DBH). Adrenaline-releasing cells alone harbor phenylethanolamine-N methyl transferase (PNMT), an enzyme that methylates noradrenaline, converting it to adrenaline. Adrenomedullin (AM) is a hormone that is highly expressed in adrenal medullary mesenchyme. To clarify mimecan expression in adrenal medulla, we performed dual-color in situ hybridization of mimecan and PNMT or TH or AM. As a result, neither PNMT (blue) expression in adrenaline-releasing cells of adrenal medulla nor TH (blue) expression in most of adrenal medulla was co-expressed with mimecan (red) (Fig. 1C-F). However, AM (blue) and mimecan (red) are co-expressed in the medullary mesenchyme ( Fig. 1G and H). In conclusion, these results revealed that, in the mouse adrenal gland, mimecan is mainly expressed in the cortex although being also detected in the medulla.
The adrenal cortex is responsible for synthesis of glucocorticoid, which are essential for survival under stress. Given the high-level of adrenocortical mimecan expression demonstrated by our previous and present studies (11,12), we presumed that mimecan may be involved in regulating functions of the HPA axis during responses to stress.

Mimecan expression in adrenal tissue was decreased after acute stress
We used insulin-induced hypoglycemia in mice as a model to investigate the expression of mimecan after acute stress (18). The test mice were fasted 12h overnight and then given insulin (3 IU/kg body weight) by IP injection. Blood glucose was subsequently measured at various time points. Blood glucose levels were all below 40 mg/dl, showing a successfully induced hypoglycemia state in stressed group ( Table 1). As anticipated, expression of StAR, which is a marker of stress, was induced by hypoglycemic stress, showing signi cant upregulation 1 h after insulin dosing and peaking at 6 h ( Fig. 2A). Interestingly, signi cant time-dependent downregulation of mimecan expression was observed (Fig. 2B). However, blood glucose had normalized 8 h after insulin administration ( Table 2), implying that reduced mimecan expression constitutes a response to stress induced by hypoglycemia rather than hypoglycemia state. To exclude the possibility that StAR up-regulation and mimecan down-regulation was induced by insulin directly rather than insulin induced hypoglycemia stress, we examined StAR and mimecan mRNA levels after insulin stimulation in the Y1 cell ( Fig. 2C and 2D). No signi cant change of StAR and mimecan mRNA levels was observed 2h and 6h after insulin stimulation. To further verify this result, we established burn trauma stress model in mice as described. In situ hybridization showed that expression levels of StAR mRNA signi cantly increased in adrenal tissues of mice after scalding (Fig. 2E), so this stress model appeared sound. Northern blot showed a signi cant downregulation of mimecan in adrenal tissue, whereas levels in mouse lung and adipose tissues were unchanged after scalding (Fig. 2F). Above results revealed that Mimecan expression in adrenal tissue was signi cantly decreased after acute stress.
Stress-related mimecan downregulation was mediated directly by ACTH rather than ACTH induced glucocorticoid secretion.
Activation of the HPA axis and the accompanying hormonal response to stress is triggered by a surge of CRH into the hypothalamic-pituitary-portal system. Elevated CRH in portal blood increases ACTH secretion in the pituitary and produces a corresponding rise in adrenal glucocorticoid secretion (19). To determine whether decreased mimecan expression after stress are mediated by increased ACTH secretion, we administrated ACTH to C57BL/6 mice at doses of 0.085 U/g at 2, 4, 6, 8, and 12 h. The animals were then sacri ced at various time points, and their adrenal glands were collected bilaterally at once, using Northern blot to sequentially assess mimecan mRNA expression in adrenal tissues of these mice. The stated levels showed signi cant time-dependent declines after ACTH dosing (Fig. 3A). Compared with controls, expression of mimecan in adrenal tissues was reduced by ∼70% at 2 h and ∼90% at 6 h post-treatment (Fig. 3A), whereas corresponding levels in lung tissue were unchanged (Fig.   3B). We also detected mimecan expression after ACTH treatment in primary cultures of mouse adrenal cells (Fig. 3C). Similarly, fresh adrenal cellular isolates cultured with 1 μM ACTH for 6 h and 12 h revealed declines in mimecan mRNA levels compared with controls (Fig. 3C). Considering the complexity of mice, we again applied this strategy to the Y1 adrenocortical cell line. The cells were treated with 10 -10 , 10 -8 , and 10 -6 M ACTH for 12 h, assessing mimecan mRNA levels by Northern blot (Fig. 3D) and real-time quantitative PCR (Fig. 3F). Mimecan expression in Y1 cells was downregulated by ACTH treatment in a dose-dependent manner. Compared with controls, expression of mimecan in the Y1 cell line was reduced by 40% at 10 -8 M and 60% at 10 -6 M after a 12-h ACTH exposure ( Fig. 3D and 3F); and mimecan mRNA levels declined in a time-dependent manner after 1 μM ACTH treatment, as shown by Northern blot (Fig.  3E) and real-time quantitative PCR (Fig. 3G).
Because our previous study found that glucocorticoid up-regulates mimecan expression in corticotroph cells (11), We also examined the effects of DEX on mimecan expression in adrenal tissues. In accordance with the results in corticotrophin cells, mRNA levels of mimecan in adrenal tissues of C57BL/6 mice was markedly upregulated in a time-dependent manner after IM injection of DEX (0.05 ug/g body weight), which peaked at 12 h and were sustained for 36 h (Fig. 4A). Likewise, mimecan expression in Y-1 cells increased signi cantly in time- (Fig. 4B and 4D) and dose-dependent manners ( Fig. 4C and 4E) after DEX exposure. Moreover, the DEX-induced increase in Y1 cellular expression of mimecan was abolished by treatment with 1 μM RU486, a GR blocker (Fig. 4C and 4E). The results seem contradictory, but it may be explained by that ACTH inhibits mimecan to prevent excessive secretion of glucocorticoid after acute stress because we found mimecan also promote glucocorticoid secretion in Fig. 5A. Moreover, the fact that ACTH lowered mimecan expression and DEX increased its expression during in vivo and in vitro studies clearly indicates that the effects of stress on the downregulation of mimecan expression in adrenal tissues are mediated by ACTH rather than glucocorticoid.

Mechanism of inhibited adrenal mimecan expression due to ACTH
Binding of ACTH to its receptor generally activates the following four signaling pathways: 1) cAMP/PKA/CREB; 2) MEK/ERK; 3) PKC; and 4) JNK. The cAMP/PKA/CREB pathway is responsible for upregulated StAR expression, which is otherwise downregulated by activated MEK/ERK, PKC, or JNK pathways. To investigate those pathways mediating the effects of ACTH on mimecan expression, Y1 cells were separately treated with speci c inhibitors of four ACTH signaling pathways (SQ22536 for cAMP, H89 for PKA, PD98059 and U0126 for ERK, GO6983 for PKC, and SP600125 for JNK) for 40 minutes prior to 1 μM ACTH stimulation. As expected, mRNA levels of StAR were signi cantly upregulated in Y1 cell lines at 6 h and then abolished after treatment with cAMP/PKA inhibitors, H89 (20 μM) or SQ22536 (200 μM) ( Supplementary Fig. 1A, 1B and 1E). However, the observed lowering of mimecan mRNA in Y1 cells induced by ACTH was not reversed after treatment with the cAMP/PKA pathway inhibitor, H89 (20 μM) or SQ22536 (200 μM) (Fig 3F, Supplementary Fig. 2A and 2B). Both the PKC inhibitor GO6983 and JNK inhibitor SP600125 did not reversed the observed lowering of mimecan mRNA in Y1 cells induced by ACTH (Fig 4F, Supplementary Fig. 2C and 2D) and the stimulation effects of ACTH on StAR (Supplementary Fig. 1C, 1D and 1E). It is therefore likely that inhibition of mimecan expression in adrenal cells due to ACTH is not mediated by the cAMP/PKA/CREB, PKC, or JNK pathway. However, the inhibitory effects of ACTH on mimecan expression in Y1 cells after 6-h stimulation by 1 μM ACTH was abolished by treatment with the ERK inhibitors, PD98059 (25 μM, 50 μM) and U0126 (10 nM, 20 nM) (Fig.  4G). Consequently, ACTH downregulation of mimecan expression in adrenal cells is likely mediated by the ERK signaling pathway.

Mimecan stimulates secretion of corticosterone in mouse adrenal cells
Having shown that mimecan is downregulated by ACTH and that corticosterone increases mimecan expression in adrenal tissues, we further investigated whether mimecan regulates ACTH and corticosterone secretion. C57BL/6 mice received IP injections of mimecan-MBP fusion protein, PBS, or MBP alone at doses of 0.1 μmol/kg body weight. Whole blood was then obtained from retroorbital spaces for ACTH and corticosterone assay (ELISA). Serum corticosterone concentrations were much higher in C57BL/6 mice receiving mimecan-MBP fusion protein at the stated dose than in recipients of PBS or MBP (Fig. 5A). Compared with controls, serum corticosterone was signi cantly elevated at 0.5 h after IP injection of mimecan-MBP fusion protein, reaching a peak at 2 h, maintaining high levels for at least 4 h, and then normalizing at 24 h (Fig. 5A). Serum levels of ACTH did not differ signi cantly among any of these treatment groups (Fig. 5B). These data suggest that mimecan-induced corticosterone secretion is not mediated by heightened levels of ACTH. Compared with MBP-treated Y1 cells, the concentration of corticosterone in culture media clearly increased after treatment with mimecan-MBP fusion protein for 24 h, as opposed 8 h at a dose of 7.64 μM (Fig. 5C). To determine the mechanism for corticosterone elevation, expression levels of key genes regulating corticosterone synthesis in adrenal tissues, such as StAR, CYP11A1, CYP11B1 and CYP21, were detected using real-time PCR. Remarkably, expression levels of these genes in the adrenal tissues of mice were substantially downregulated 4 h after IP injection of mimecan-MBP fusion protein (Fig. 5D-G). Therefore, the observed mimecan-induced rise in corticosterone concentrations of mouse serum or Y1-cell culture media were probably due to intensi ed corticosterone secretion by stimulated adrenal cortical cells rather than increased corticosterone synthesis.
Mimecan de ciency disturbs stress-free diurnal rhythms of corticosterone secretion, enhancing HPA activation in a mimecan knockout mouse restraint model We further generated mimecan knockout mice to delineate the physiologic roles of mimecan (12). In stress-free states, serum corticosterone and ACTH levels ( Fig. 6A and 6B) and mRNA expression levels of genes involved in corticosterone synthesis ( Supplementary Fig. 3A-D) were similar for knockout mice and wild-type litter mates. However, the diurnal rhythm of corticosterone secretion in knockout vs wild-type mice was disturbed signi cantly, as shown by serial determinations of serum corticosterone collected at 8:00, 12:00, 16:00, 20:00, 0:00, and 4:00 hours in an unstressed state (Fig. 6C). In wild-type mice, these levels peaked in the evening (20:00 hours) and reached a nadir in the morning, re ecting non-stressful circadian secretion of cortisol (Fig. 6C). However, no clear peaks or troughs of serum corticosterone were encountered in counterpart knockout mice, only a attened pattern (Fig. 6C). To explore the responses of mimecan knockout mice to ACTH and glucocorticoid, these mice were received IP injections of ACTH (0.085 IU/g body weight) and IM injections of DEX (0.05 ug/g body weight). Although serum corticosterone levels in knockout and wild-type mice after ACTH stimulation did not differ signi cantly (Fig. 6D), serum corticosterone levels were lower in knockout vs wild-type mice following IM DEX injection for 5h to 32h (Fig. 6E), which may due to the absence of corticosterone boosting effects of mimecan in knockout mice. Meanwhile, serum ACTH levels remained comparable in both groups of mice (Fig. 6F), and mRNA expression levels of genes encoding corticosterone synthesis enzymes also showed no difference between knockout and wild-type mice after IM DEX injection. (Supplementary Fig. 3E-H).
We also investigated whether genetic mimecan de ciency in uenced pituitary and adrenal gland function in response to stress. Male knockout mice and wild-type litter mates were restrained for 60 min to in ict physical stress. We collected blood through tail cuts at the time of release and 60 min later, measuring serum corticosterone by ELISA. Restraint of knockout mice for 60 min induced a signi cantly greater increase in serum corticosterone compared with levels in wild-type litter mates (Fig. 6G). Moreover, the observed difference in serum corticosterone concentrations between knockout and wild-type mice was lost 60 min after release (Fig. 6G). These ndings indicate a striking excess of the response to stress in knockout mice.

Discussion
Although implicated in many biologic processes, the physiologic functions of mimecan remain elusive. It is most extensively researched in keratocytes of the bovine cornea, where it likely helps maintain matrix homeostasis (20,21). Apart from its important roles in the regulating construction of extracellular matrix, the expression de ciency or down regulation of mimecan has been indicated in several kinds of tumors (22)(23)(24). A prior study of ours aimed at lung cancer has shown the considerable promise of mimecan as a biomarker, helping to distinguish non-small-cell from small-cell lung cancer pathologic types (25). Expression of mimecan mRNA and protein were also observed in the human anterior pituitary glands. Luciferase reporter analysis and electrophoretic mobility shift assays indicate that pituitary transcription factor-1 (Pit-1), expressed solely in the pituitary, is capable of activating the human mimecan promoter via Pit-1 response-element sites (18). Most recently, we have discovered that as a newly recognized adipokine, mimecan can inhibit food intake by inducing IL-1β and IL-6 expression in the hypothalamus (10). This body of evidence attests to the diverse actions of mimecan in many physiologic and pathologic processes.
According to the present ndings, mimecan is primarily expressed in adrenal cortex. We found that its expression in adrenal tissue was signi cantly downregulated under stress in our mouse model, signifying a role in stress response. We also determined that expression of mimecan in adrenal cells may be downregulated by ACTH and upregulated by DEX. Although both dramatic increases in levels of secreted ACTH and corticosterone were apparent in mice after stress, the data herein clearly illustrate that stressinduced downregulation of mimecan in adrenal tissues is mediated by ACTH, rather than corticosterone.
Cortisol is the terminal effector of the HPA axis, and cortisol secretion is stimulated as a result of HPA activation in response to acute stress. It is generally acknowledged that cortisol secreted by adrenal cortex is directly tied to pituitary ACTH release. In recent years, however, there is mounting evidence that many adrenomedullary secretory products, such as adrenomedullin, catecholamines, 5-HT, neuropeptides, and growth factors, take part in regulating adrenocortical steroidogenesis via non-ACTH-dependent pathways (26)(27)(28). Furthermore, cytokines produced by adrenal cells themselves are capable of directly in uencing adrenocortical function. They may stimulate steroidogenesis (similar to IL-1, IL-2, and IL-6) or exert regulatory effects on adrenal growth (as does TNFα or interferon-γ) (29)(30)(31)(32). It is a proven fact that cortisol release is self-regulated through a CNS negative feedback loop involving receptors in pituitary, PVN, and other brain structures(33). However, little is known of similar synthesis/secretion autoregulation via intra-adrenal mechanisms.
In the present study, based on in vivo and in vitro studies, we found that mimecan stimulates corticosterone secretion from adrenal tissues in a short time response, yet there were no in vivo effects on serum levels of ACTH. Additionally, mimecan expression may be upregulated by corticosterone, as shown by DEX administration studies (in vitro and in vivo), and this effect vanished after treating Y-1 adrenal cells with RU486 (a GR inhibitor), which is in accord with our previous results that mimecan expressed in pituitary corticotrophin cells is increased by glucocorticoid and that its upregulation may be mediated through classic GR pathways (11). Our ndings therefore suggest that under basal conditions, mimecan may stimulate glucocorticoid release from adrenal cells as a means of positive feedback (Fig. 6H), although this premise is still tentative. Collectively, our data support the concept of ACTH-independent glucocorticoid regulation, permitting self-regulation of synthesis and secretion within the adrenal gland owing to upregulation of intra-adrenal mimecan expression.
Another interesting facet of this study was that mimecan de ciency in mice fosters disturbances in diurnal rhythms of corticosterone secretion and stronger responses to restraint stress, despite the similar serum ACTH and corticosterone concentrations seen in wild-type and mimecan knockout mice under basal conditions. Consequently, it is tempting to presume that mimecan plays a vital intra-adrenal role, helping to maintain glucocorticoid homeostasis in mice after acute stress. In basal states of moderatelevel ACTH release, glucocorticoid promotes mimecan expression in adrenal tissues through GRs, in turn prompting secretion of glucocorticoid to sustain appropriate levels. Thus, a positive intra-adrenal feedback loop seems to be at work. However, stress-related increases in ACTH release due to HPA axis over-activation will then culminate in downregulation of intra-adrenal mimecan expression, attenuating its effects on glucocorticoid secretion and avoiding glucocorticoid excess. In mimecan knockout mice, this negative feedback effect during acute stress is lost, explaining their comparatively stronger responses to stress.
In conclusion, we have demonstrated for the rst time the critical role that mimecan likely plays in regulating the HPA axis. In the absence of mimecan effects, we observed disturbed circadian rhythms of corticosterone secretion in C57BL/6 mice and abnormal secretion of corticosterone in response to stress or exogenous steroid stimulation. Although ACTH may inhibit mimecan expression in adrenal tissues, any mechanisms of mimecan control are still unknown. Future studies investigating the impact of mimecan on corticosterone secretion are clearly warranted.

Declarations
Ethics approval and consent to participate The animal study was performed in accordance with the approved protocol by the Animal Care and Use Committee in the Ninth Hospital A liated to Shanghai Jiaotong University School of Medicine.

Consent for publication
Not applicable.
Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest.    Mimecan gene expression in Y-1 cells was analyzed by Northern blot. The relative mimecan mRNA levels were normalized to GAPDH mRNA expression. Data information: *p<0.05, **p<0.01, ***p<0.001 for DEX treatment vs control.