Serum RARRES2 Is a Prognostic Marker in Patients With Adrenocortical Carcinoma

Yi Liu-Chittenden, Dhaval Patel, Kelli Gaskins, Thomas J. Giordano, Guillaume Assie, Jerome Bertherat, and Electron Kebebew

Endocrine Oncology Branch (Y.L .- C., D.P., K.G., E.K.), National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892; Department of Pathology (T.J.G.), University of Michigan, Ann Arbor, Michigan 48109; Institut Cochin (G.A., J.B.), Inserm Unité 1016, Centre National de la Recherche Scientifique Unité Mixte de Recherche 8104, Descartes University, 75014, Paris, France; and Department of Endocrinology (G.A., J.B.), Reference Center for Rare Adrenal Diseases, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, 75014, Paris, France

Context: Retinoic acid receptor responder protein 2 (RARRES2) is a small secreted protein involved in multiple cancers, including adrenocortical carcinoma (ACC). However, discordant tumor and serum RARRES2 levels have been reported in various cancers. The etiology of this discordance is unknown and has not been studied in pair-matched tumor and serum samples.

Objective: To determine tissue and serum RARRES2 levels in patients with adrenocortical neoplasm and to elucidate the prognostic implications of RARRES2 levels.

Design, Settings, and Patients: Tissue and serum RARRES2 levels were analyzed. A pair-matched analysis was performed to examine tissue and serum RARRES2 from 51 patients with benign ad- renocortical tumors and 18 patients with ACC. Overall survival was analyzed based on RARRES2 expression. A mouse xenograft model was used to determine the source of serum RARRES2.

Results: Patients with ACC had decreased tumor RARRES2 gene expression (P < . 0001) and in- creased serum RARRES2 levels (P < . 005) as compared with patients with benign adrenocortical tumors. Higher serum RARRES2 levels were associated with improved overall survival (P = . 0227). A mouse xenograft model demonstrated that higher tissue RARRES2 expression was associated with higher RARRES2 secretion in the serum and that there was an intrinsic mechanism in main- taining serum RARRES2 homeostasis.

Conclusions: Serum and tissue RARRES2 expression levels are paradoxical in patients with ACC. The elevated RARRES2 in patient serum is unlikely to be secreted from tumor cells. Serum RARRES2 may be used as a novel prognostic marker for ACC. (J Clin Endocrinol Metab 101: 3345-3352, 2016)

I Incidentally detected adrenal incidentalomas have been increasing in prevalence with the advent of the com- puted tomography scan (1). These incidentalomas are common compared with adrenocortical carcinoma (ACC), but many patients undergo an adrenalectomy to exclude the diagnosis of ACC. ACC, a rare malignancy with an annual incidence of 0.7-2.0 cases per million peo- ple (2-4), has a poor prognosis. The 5-year overall sur- vival rate ranges from 32% to 42% but is heterogeneous

(2, 5). Furthermore, even after complete tumor resection, over half of the patients develop recurrent disease. Many patients undergo chemotherapy, which consists of a reg- imen, including mitotane plus combination chemotherapy with etoposide, doxorubicin, and cisplatin. Unfortu- nately, this regimen has limited impact on overall survival and low response rates (6, 7). Understanding the mecha- nism behind disease progression and identifying biomark- ers for diagnosis and prognosis are of great importance in

Abbreviations: ACC, adrenocortical carcinoma; BMI, body mass index; GAPDH, glyc- eraldehyde-3-phosphate dehydrogenase; RARRES2, retinoic acid receptor responder protein 2.

selecting patients who may benefit from intervention and/or adjuvant therapy. Furthermore, patients with a poor prognosis may be appropriately counseled and re- ferred for clinical trials.

Retinoic acid receptor responder protein 2 (RARRES2) (also known as chemerin) is a small secreted protein that was first identified as a ligand for the orphan G protein- coupled receptor chemokine-like receptor 1 (8). To date, RARRES2 is mainly recognized for its 2 major roles as a chemoattractant and an adipokine (9). As a chemoattrac- tant, it promotes chemotaxis of chemokine-like receptor 1-expressing immune cells, including immature dendritic cells, macrophages, and natural killer cells, to lymphoid organs and sites of injury (8, 10-12). As an adipokine, it regulates adipogenesis and adipocyte metabolism (13, 14). Several studies have shown the potential involvement of RARRES2 in cancer. RARRES2 is transcriptionally down-regulated in ACC (15-17). Its expression is also reported to be reduced in melanoma, skin squamous cell carcinoma, hepatocellular carcinoma, and in lung, breast, colon, and metastatic prostate cancers (18-20). However, RARRES2 protein levels in peripheral blood were re- ported to be significantly higher in patients with lung can- cer than in healthy volunteers (21). Similarly, increased serum RARRES2 levels were detected in patients with gastric cancer as compared with healthy subjects (22, 23). To our knowledge, no studies have been conducted to determine RARRES2 expression level concomitantly in tissue and serum with pair-matched cancer patient sam- ples. The mechanism of this paradoxical tissue and serum RARRES2 expression is also unknown.

In this study, we performed a pairwise comparison of tissue and serum RARRES2 levels in patients with adre- nocortical tumors to determine the diagnostic and prog- nostic value of RARRES2 as a biomarker, and to elucidate the source of serum RARRES2 levels.

Materials and Methods

Human tissue samples

Twenty-one normal adrenal glands were obtained at the time of nephrectomy for organ donation. Fifty-one benign adreno- cortical tumor samples and 26 ACC samples (primary or local recurrence) were procured immediately after surgical resection. All tissue samples were snap frozen and stored at -80℃. Tu- mors were classified as benign when the Weiss criteria scores were less than 3, and there was no evidence of recurrence or metastasis during follow-up (mean follow-up, 3.2 y). Tumors were classified as ACC when the Weiss criteria scores were more than or equal to 3 (Table 1).

Human serum samples

Twenty-one fasting serum samples were obtained from healthy volunteers. Fifty-three fasting preoperative serum sam- ples were obtained on the day of surgery from patients with benign adrenocortical tumors with Weiss criteria scores of less than 3 and no evidence of recurrence or metastasis during follow- up. Twenty fasting preoperative serum samples were obtained on the day of surgery from patients with primary, locoregional re- currence and distant metastatic ACC with Weiss criteria scores of more than or equal to 3. Serum samples were collected and stored at -80°℃ (Tables 2 and Table 3). All patients gave written informed consent for sample collection.

RNA isolation, cDNA synthesis, and TaqMan real-time quantitative PCR

Total RNA was isolated using the RNA/DNA/Protein Puri- fication Plus kit (Norgen Biotek Corp), according to the manu- facturer’s protocol. The RNA was reverse transcribed using the High Capacity cDNA Reverse Transcription kit (Applied Bio- systems). TaqMan real-time quantitative polymerase chain re- action was performed on 7900HT fast real-time PCR systems (Applied Biosystems). The reaction was prepared by mixing cDNA, TaqMan 2X universal PCR master mix and TaqMan gene expression assays (Applied Biosystems). The gene expression assays used were: RARRES2 (Hs00161209_g1) and glyceraldehyde-3- phosphate dehydrogenase (GAPDH) (Hs99999905_m1).

Enzyme-linked immunosorbent assay

Human Chemerin Quantikine ELISA kit (limit of quantifi- cation, 78 pg/mL) or mouse Chemerin Quantikine ELISA kit

Table 1. Clinical Features of Tissue Samples Used in Tissue RARRES2 Expression Assessment
Benign Adrenocortical
TumorACC
Number of patients5126
Age (average ± SD)47.3 ± 15.648.0 ± 13.7
Sex (female/male)32/1917/9
Type of tumorPrimary8 primary, 3 locoregional recurrence without distant metastasis, 15 locoregional recurrence with distant metastasis
Syndromeª
Adrenal hypercortisolism2711
Subclinical adrenal hypercortisolism20
Primary hyperaldosteronism191
Nonfunctioning314

a Functional status at initial presentation.

Table 2. Clinical Features of Tissue Samples Used in Serum RARRES2 Expression Assessment
Healthy ControlsBenign Adrenocortical TumorACC
Number of patients215320
Age (average ± SD)45.8 ± 12.947.8 ± 15.650.2 ± 10.5
Sex (female/male)10/1132/2112/8
Type of tumorPrimary1 primary, 3 locoregional recurrence without distant metastasis, 16 locoregional recurrence with distant metastasis
Syndromeª
Adrenal hypercortisolism276
Subclinical adrenal20
hypercortisolism
Primary hyperaldosteronism211
Nonfunctioning313

a Functional status at initial presentation.

(limit of quantification, 47 pg/mL) (R&D Systems) were used to measure human or mouse RARRES2 levels in human and mouse serum samples. All samples were measured in duplicate at 1:100 dilutions.

Cell lines and animals

Human ACC cell lines NCI-H295R (H295R) were purchased from ATCC and authenticated by short-tandem repeat profiling. Cells were cultured in DMEM supplemented with 1% insulin transferrin selenium and 2.5% NuSerum I (Corning) and main- tained in a standard humidified incubator with 5% CO2 at 37ºC. RARRES2-overexpression plasmid was generated by cloning

Table 3. Clinical Characteristics of Patients With ACC for Survival Analysis
Patient Characteristics
Sex
Male, n (%)8 (40%)
Female, n (%)12 (60%)
Age at diagnosis, median (range)51 (32-68)
Age at operation, median (range)55 (34-72)
BMI, mean (range)26.5 (17.7-35.5)
Syndromeª
Hyperaldosteronism1
Hypercortisolism6
Nonfunctional13
Extent of disease
Primary1 (5%)
Locoregional disease, n (%)3 (15%)
Distant mets, n (%)16 (80%)
Previous treatment
Surgery, n (%)b16 (80%)
Chemotherapy, n (%)"17 (85%)
Ki67 (%)d
<5%, n (%)2 (12.5%)
>20%, n (%)14 (87.5%)

a Functional status at initial presentation.

b Previous surgeries include primary resection of ACC, metastectomy, radiofrequency ablations, and debulking procedures.

” Patients received chemotherapeutic regimens, including combinations of mitotane, etoposide, doxorubicin, cisplatin, protease inhibitors, abraxane, tariquidar, imatinib, streptozocin, or cetuximab.

d Ki67% for patients with available data.

RARRES2 cDNA from pCMV-XL5-RARRES2 (Origene) into the pcDNA4/Myc-His A vector (Life Technologies) to generate pcDNA4-RARRES2 so that only the untagged protein was ex- pressed. Polyclonal stable cell lines overexpressing RARRES2 or an empty vector were generated by transfecting H295R cells with Bg/II (New England Biolabs) linearized pcDNA4-RARRES2 plasmid or the pcDNA4/Myc-His A vector, followed by selection with 1-ug/mL Zeocin (Life Technologies). Immunodeficient NOD.Cg-Prkdcscid Il2rgtm1Wil/SzJ (NOD scid-y) mice were pur- chased from The Jackson Laboratory and were maintained and bred under the guidelines of the National Institutes of Health Animal Research Advisory Committee. Mice littermates that were 3-4 months old were randomly assigned to 2 groups with 7 mice in each group. Each mouse received sc flank injections bilaterally with 100 µL of a 1:1 mixture of either medium or 4 X 106 cells with Matrigel (Corning, Inc). Tumor sizes were mea- sured and recorded once a week. Mice were euthanized when tumor sizes reached 2 cm in diameter. Whole-blood samples were collected twice: before injection from the facial vein and immediately after euthanization by cardiac puncture. Serum was obtained by centrifugation of blood samples for 15 minutes at 4000 rpm.

Western blot analysis

Total protein lysates were subjected to SDS-PAGE gel elec- trophoresis, transferred to polyvinylidene difluoride mem- branes, and immunostained with primary antibodies overnight at 4℃. The primary antibodies used were the RARRES2 anti- body (1:1000, ab72965; Abcam) and GAPDH antibody (1: 5000, sc-47724; Santa Cruz Biotechnology, Inc). The secondary antibody used was goat antimouse IgG-horseradish peroxidase (1:5000, sc-2005; Santa Cruz Biotechnology, Inc). Proteins were detected with SuperSignal West Pico or West Femto Chemilu- minescent Substrate (Thermo Fisher Scientific).

Statistical analysis

Mann Whitney tests were used to compare RARRES2 levels between different groups. The primary outcome variable ana- lyzed was overall survival. Overall survival was defined as the time from preoperative blood samples drawn to disease-related mortality or last follow-up, which was censored for those pa- tients who underwent repeated abdominal surgery. The associ- ation between human serum RARRES2 levels and survival was

assessed using the Kaplan-Meier method, with statistical differ- ences determined by log-rank test. P < . 05 was considered sta- tistically significant. All calculations were performed using GraphPad Software 6.0.

Results

RARRES2 gene expression level is down-regulated in ACC tissue samples but is not associated with prognosis

RARRES2 gene expression levels were measured in 21 normal adrenocortical tissues, 51 benign adrenocortical tu- mors and 26 ACC samples. Significant down-regulation of RARRES2 expression was observed in ACC samples when

compared with benign or normal samples (P <. 0001 in both comparisons); it was also observed in benign samples when compared with normal adrenocortical tissues (P < . 0001) (Figure 1A). These results are consistent with those from our previous genome-wide gene expression study in another in- dependent cohort, in which RARRES2 was down-regulated in primary ACC samples (n = 5) as compared with benign adrenocortical tissues (n = 74) (15). Similar results were also obtained when we analyzed 2 publicly available adrenocor- tical tissue databases (NCBI-GEO, accession number GSE10927; and EMBL-EBI, accession number E-TABM- 311) (Figure 1, B and C) (24, 25). To determine whether tissue RARRES2 expression is associated with prognosis, we analyzed whether RARRES2 expression levels in ACC tissue

Figure 1. The tissue RARRES2 gene expression level is down-regulated in ACC but is not associated with patient prognosis. A, TaqMan real-time quantitative PCR reveals down-regulation of the RARRES2 gene expression level in our cohort of ACC samples as compared with benign adrenocortical tumors. Human RARRES2 expression is presented as % gene expression (= 2 -A(Ct RARRES2 - Ct GAPDH) × 100). B, RARRES2 expression as analyzed in NCBI-GEO database GSE10927. Data are presented as relative ratio to normal tissue. C, RARRES2 expression as analyzed in European Bioinformatics Institute database E-TABM-311. Data are presented as relative ratio to normal tissue. The tumor RARRES2 gene expression level is not associated with prognosis in 4 independent datasets, including our current cohort (D), patient cohort reported in GSE10927 (E), patient cohort reported in E-TABM-311 (F), and patient cohort reported in Demeure et al (26) (G). High and low RARRES2 groups were defined by median tissue gene expression. ns, not significant.

A

P <0.0001

B

P< 0.0001

C

P < 0.01

P <0.0001

P < 0.05

P < 0.0001

ns

P < 0.0001

1.5

2.5

3000

% Gene expression

:

Ralative ratio

2.0-

Ralative ratio

2000

1.0

P <0.0001

1.5

1.0-

1000-

0.5

0.5

0.0

0

0.0

Normal (n=21)

Benign (n=51)

ACC (n=26)

Normal (n=10)

Adenoma (n=22)

ACC (n=33)

Normal (n=4)

Adenoma (n=58)

ACC (n=34)

D

100

- High Tissue RARRES2 (n=9)

E

—. Low Tissue RARRES2 (n=9)

100

- High Tissue RARRES2 (n=12)

Percent survival

80

P = 0.3455

Percent survival

80

—. Low Tissue RARRES2 (n=12)

60-

60

P = 0.1556

40-

40

20

20

0

0

0

500

1000

1500

0

5

10

15

20

Time (days)

Time (Years)

F

G

100

- High Tissue RARRES2 (n=17)

100

- High tissue RARRES2 (n=11)

Percent survival

80-

—. Low Tissue RARRES2 (n=17)

Percent survival

80

—. Low tissue RARRES2 (n=11)

60

P= 0.9139

60

P = 0.9902

40

40

20

20

0

0

0

50

100

150

0

5

10

15

20

Time (Months)

Time (Years)

Figure 2. The serum RARRES2 level is up-regulated in patients with ACC and is associated with prognosis. A, The serum RARRES2 level was significantly elevated in patients with ACC as compared with healthy controls and patients with benign adrenocortical tumors. B, Kaplan-Meier analysis for the serum RARRES2 level showed a significant difference in survival based on serum RARRES2 levels (P = . 0227). The median survival for the high serum RARRES2 group is 1419 days, and the median survival for the low-serum RARRES2 group is 778 days. The median follow-up is 2.3 years. High and low RARRES2 groups were defined by median serum RARRES2 level. ns, not significant.

A

P < 0.001

B

P < 0.005

100

- High Serum RARRES2 (n=10)

ns

250

80

Serum RARRES2 (ng/ml)

Percent survival

—. Low Serum RARRES2 (n=10) P = 0.0227

200-

60

150

40

100

20

50

0

0

500

1000

1500

Time (days)

0

Normal (n=21)

Benign (n=53)

ACC (n=20)

samples were associated with patient overall survival in our current patient cohort and patient cohorts depicted in GSE10927, E-TABM-311, and Demeure et al (26). No sig- nificant association was found between tissue RARRES2 ex- pression and patient overall survival (Figure 1, D-G).

Serum RARRES2 protein level is elevated in patients with ACC and is associated with prognosis

Serum RARRES2 protein levels were measured in se- rum samples collected from 21 healthy volunteers, 53 pa- tients with benign adrenocortical tumors, and 20 patients with ACC. Among these, serum samples from 51 patients with benign tumors and 18 patients with ACC were paired with the tissues analyzed for RARRES2 gene expression. Serum RARRES2 levels were significantly elevated in pa- tients with ACC compared with patients with benign tu- mors (P <. 005) and healthy controls (P <. 001), whereas no significant difference was observed between healthy controls and patients with benign tumors (Figure 2A). Se- rum RARRES2 levels were not significantly different based on tumor functional status, disease burden, gender, or maximum standardized uptake value, as determined by fluorodeoxyglucose positron emission tomography/com- puted tomography (data not shown). However, survival analysis based on median serum RARRES2 levels (105 ng/ml) showed that higher RARRES2 serum levels were associated with longer overall survival (P = . 0227; median survival of 1419 days for high serum RARRES2 group vs 778 days for low-serum RARRES2 group; median follow- up, 2.3 y) (Figure 2B).

Elevated serum RARRES2 levels are not from tumor cell secretion

Given the low tissue RARRES2 levels yet high serum RARRES2 levels in patients with ACC, we sought to un-

derstand the reason for this apparent paradox. Pair- matched tissue and serum samples from 51 patients with benign adrenocortical tumors and 18 patients with ACC were analyzed for RARRES2 expression, which con- firmed the existence of this paradox within the same pa- tient cohort (Figure 3A). However, no significant corre- lation between tissue and serum RARRES2 levels was observed (r = - 0.1761, P = . 1478) (Figure 3B). To de- termine whether the elevated serum RARRES2 levels came from active secretion of RARRES2 from ACC tissues, we conducted a mouse xenograft experiment. Immunodefi- cient NOD scid-y mice were randomly assigned to 2 groups: one group received injections of H295R cells sta- bly transfected with a plasmid expressing RARRES2 (RAR), and the second group received injections of H295R cells stably transfected with an empty vector (Fig- ure 3C). The mice were monitored for 8 weeks before reaching the endpoint of the study. All xenografts devel- oped into tumors except for 1 xenograft in the RARRES2 group. In this experimental model, we can distinguish the origin of RARRES2 in the serum: human RARRES2 should come from the xenografted human cells expressing human RARRES2, whereas mouse RARRES2 should come from the mouse host environment. As expected, the human RARRES2 level in the serum was high in the RAR group after the xenograft experiment (Figure 3D). The amount of secreted human RARRES2 was significantly correlated with tumor weight (r = 0.9883, P < . 0001), indicating that more cells with RARRES2 expression led to increased RARRES2 secretion and detection in the se- rum (Figure 3E). The vector group members with minimal basal human RARRES2 expression developed tumors with minimal detection of human RARRES2 in serum (Figure 3D). Before the xenograft experiment, no signifi- cant difference in mouse serum RARRES2 levels was ob-

Figure 3. Elevated serum RARRES2 levels observed in patients do not come from tumors expressing low RARRES2. A, Pair-matched analysis for tissue RARRES2 (left) and serum RARRES2 (right) levels from 51 patients with benign adrenocortical tumors and 18 patients with ACC. B, Tissue RARRES2 expression has no significant correlation with the serum RARRES2 level from the same patients. r, correlation coefficient. Pair-matched tissue and serum RARRES2 levels from 51 benign adrenocortical tumors and 18 ACC samples were examined. % gene expression = 2 -A(Ct RARRES2 - Ct GAPDH) × 100. C, Representative Western blot showing the difference in the RARRES2 protein level in stable H295R cells overexpressing RARRES2 (RAR) as compared with stable H295R cells overexpressing an empty vector. The stable cell lines were used in mouse xenografts. D, The human RARRES2 level increased significantly in mouse serum when the mice were xenografted with stable RARRES2 overexpressing H295R cells. E, The human RARRES2 level in mouse serum correlated significantly with tumor weight. F, Mouse serum RARRES2 levels before the xenograft experiment. No significant difference was observed. G, Mouse serum RARRES2 levels after the xenograft experiment. Mice xenografted with RARRES2-overexpressing ACC cells showed significantly decreased mouse RARRES2 levels. H, Total (mouse and human) RARRES2 levels in mouse serum after the xenograft experiment. No significant difference was observed between the groups.

A

B

600

P < 0.0001

Serum RARRES2 (ng/ml)

250

P < 0.005

250

r = - 0.1761

200

P = 0.1478

% Gene expression

Serum RARRES2 (ng/mL)

200

400

150

150

200

100

100

50

0

50

0

Benign (n=51)

ACC (n=18)

0

Benign (n=51)

ACC (n=18)

0

200

400

600

Tissue RARRES2 (% gene expression)

C

D

After xenograft

E

RAR

Human RARRES2 (ng/mL)

P < 0.001

40

Vector

Human RARRES2 (ng/ml)

r = 0.9883

35

P < 0.0001

30

25

RARRES2

15

20

5

10.

GAPDH

8:3

0

Vector (n=7)

RAR (n=7)

0.0

0.5

1.0

1.5

2.0

2.5

Tumor weight (g)

F

Before xenograft

G

After xenograft

H

After xenograft

Mouse RARRES2 (ng/ml)

250

Mouse RARRES2 (ng/ml)

250

P < 0.05

Total RARRES2 (ng/ml)

250

200

200

200

F

È

150

150

150

100

100

100

50

50

50

0

Vector (n=7)

RAR (n=7)

0

Vector (n=7)

RAR (n=7)

0

Vector (n=7)

RAR (n=7)

served between the groups (Figure 3F). After the xenograft experiment, mouse serum RARRES2 levels in the RAR group significantly decreased compared with the vector group (Figure 3G), but total serum RARRES2 levels re- mained constant in both groups (Figure 3H).

Discussion

In the present study, tissue and serum RARRES2 levels were examined in healthy controls and patients with benign and malignant adrenocortical tumors. A pair-matched analysis

was performed for tissue and serum RARRES2 levels from patients with benign adrenocortical tumors and ACC. High serum levels of RARRES2 was found to indicate improved survival. To our knowledge, this is the first study to examine tissue and serum RARRES2 levels concomitantly in the same cohort for any cancer.

Our data show a paradoxical phenomenon where pa- tients with ACC have lower RARRES2 in the tumor tissue but higher RARRES2 in the serum as compared with pa- tients with benign tumors and healthy controls. These findings are consistent with published reports in patients

with lung cancer (20, 21), although paired tumor tissue and serum analysis for RARRES2 was not conducted in these studies. Using an immunodeficient mouse model xe- nografted with human ACC cells, we showed that higher tumor RARRES2 levels correlated with higher RARRES2 secretion into the serum. These results provide evidence against the possibility that ACCs with low tissue RARRES2 expression will have high RARRES2 secretion into the serum. Therefore, high serum RARRES2 levels may be coming from the host environment instead of tu- mor cells. In fact, in our study, we showed that when tumors were actively secreting RARRES2 into the host environment, host-originated RARRES2 decreased ac- cordingly so that total serum RARRES2 level remained unchanged. This suggests that there might be an intrinsic mechanism in maintaining homeostasis in the serum RARRES2 level. Thus, it is possible that in patients with ACC which have low RARRES2 tissue expression hence less RARRES2 secretion into the serum, the host environ- ment will release more RARRES2 into the serum in order to keep the balance. The source of serum RARRES2 is unknown. RARRES2 was shown to be strongly expressed in pancreas, liver and adrenal gland (27), which may ac- count for the paradoxical findings. It has been reported that RARRES2 is strongly expressed in the white adipose tissue and that the circulating RARRES2 level is associated with key features of obesity and metabolic syndrome, such as body mass index (BMI), plasma triglycerides, and blood pressure (13, 14). However, it is unlikely that the serum RARRES2 elevation observed in patients with ACC was caused by obesity, as no studies have suggested a link be- tween ACC and obesity. Furthermore, there was no asso- ciation between BMI and serum RARRES2 levels in our cohort, and the BMI was not significantly different be- tween patients with benign adrenocortical tumors and ACC (data not shown). Research has also suggested that inflammation can stimulate RARRES2 release from adi- pose tissues into the serum (28). The proinflammatory cytokine TNFa has been shown to stimulate RARRES2 release from adipocytes, whereas anti-TNF& monoclonal antibody therapy can reduce RARRES2 secretion in rheu- matoid arthritis and chronic plaque psoriasis (29, 30). However, because ACC is not known to be an immuno- genic cancer, further investigations are required to deter- mine the cellular source of RARRES2 in the serum of pa- tients with ACC.

An analysis of serum RARRES2 levels determined that these levels were a prognostic marker of overall survival in patients with ACC. Higher levels of serum RARRES2 were associated with better survival rates. When examin- ing this cohort in detail, most patients had recurrent ACC. These patients with recurrent ACC could be prognosti-

cally stratified based on serum RARRES2 levels when de- ciding on therapeutic management. Erdogan et al (31) have shown that in reoperative surgery for recurrent ACC, patients with a disease-free interval of greater than 6 months and microscopic and gross complete resection had longer survival rates. Thus, patients with high levels of RARRES2 in combination with long disease-free intervals may be ideal candidates for surgical resection if their dis- ease is resectable. Patients with lower levels may benefit from chemotherapy or clinical trials.

One interesting result from our study is that serum RARRES2 levels were higher in patients with ACC as compared with patients with benign lesions or healthy controls; however, higher serum RARRES2 levels in pa- tients with ACC were associated with better overall sur- vival. We believe that 2 major factors might contribute to this seemingly contradictory phenomenon. First, the higher serum RARRES2 level in patients with ACC does not come from tumors but from the host environment. Second, RARRES2 can act as a chemoattractant and recruit immune cells to site of inflammation (8, 10-12). In fact, in melanomas, it was shown that RARRES2 can suppress tumor growth by recruiting natural killer cells to exhibit antitumor defense (20). Thus, it is possible that ACC tissues express less RARRES2 as a mean of immune evasion (32), whereas the host environment can counteract this attempt by producing more RARRES2 in the serum. As a result, serum RARRES2 levels are higher in patients with ACC; and higher RARRES2 levels, possibly resulting from better host response, are associated with better overall survival.

Because ACC is a rare malignancy (2, 4), the small sam- ple size is a major limitation of this study. Further studies with larger patient cohorts, including more patients with primary ACC, would be beneficial in determining the value of serum RARRES2 as a diagnostic marker for dif- ferentiating between benign and malignant adrenocortical tumors. Whether this data can be generalized to patients with primary adrenal tumors without metastatic disease or patients undergoing initial surgery without metastatic disease needs to be determined and is an exciting avenue to pursue in light of the current findings. It would also be interesting to know whether serum RARRES2 levels change with disease progression and or response to treat- ment on serial blood samples. Unfortunately, we did not have serial serum samples to test this hypothesis but this would be interesting to evaluate in future studies.

In summary, we found that patients with ACC have low RARRES2 gene expression in the tumor but high RARRES2 protein levels in the serum. Serum level, but not tissue RARRES2 level, can serve as a potential prognostic marker for ACC. Our data suggests that the elevated se-

rum RARRES2 level does not originate from ACC tumor tissue. Understanding the mechanism behind the discor- dant tissue and serum RARRES2 expression, and identi- fying the source of serum RARRES2 and discovering the physiological role of RARRES2 in ACC will provide more insight into to the role of RARRES2 in ACC.

Acknowledgments

Address all correspondence and requests for reprints to: Electron Kebebew, MD, Endocrine Oncology Branch, National Cancer Institute, National Institutes of Health, 10 Center Drive, Room 4-5952, Bethesda, MD 20892. E-mail: kebebewe@mail.nih.gov.

This work was supported by the Intramural Research Pro- gram of the Center for Cancer Research, National Cancer Insti- tute, National Institutes of Health (1ZIABC01127506).

Disclosure Summary: The authors have nothing to disclose.

References

1. Toniato A, Merante-Boschin I, Opocher G, Pelizzo MR, Schiavi F, Ballotta E. Surgical versus conservative management for subclinical Cushing syndrome in adrenal incidentalomas: a prospective ran- domized study. Ann Surg. 2009;249:388-391.

2. Bilimoria KY, Shen WT, Elaraj D, et al. Adrenocortical carcinoma in the United States: treatment utilization and prognostic factors. Cancer. 2008;113:3130-3136.

3. Kamenický P, Droumaguet C, Salenave S, et al. Mitotane, metyrap- one, and ketoconazole combination therapy as an alternative to rescue adrenalectomy for severe ACTH-dependent Cushing’s syn- drome. J Clin Endocrinol Metab. 2011;96:2796-2804.

4. Fassnacht M, Libé R, Kroiss M, Allolio B. Adrenocortical carci- noma: a clinician’s update. Nat Rev Endocrinol. 2011;7:323-335.

5. Kerkhofs TM, Verhoeven RH, Van der Zwan JM, et al. Adreno- cortical carcinoma: a population-based study on incidence and sur- vival in the Netherlands since 1993. Eur J Cancer. 2013;49:2579- 2586.

6. Fassnacht M, Allolio B. Clinical management of adrenocortical car- cinoma. Best Pract Res Clin Endocrinol Metab. 2009;23:273-289.

7. Fassnacht M, Terzolo M, Allolio B, et al. Combination chemother- apy in advanced adrenocortical carcinoma. N Engl J Med. 2012; 366:2189-2197.

8. Wittamer V, Franssen JD, Vulcano M, et al. Specific recruitment of antigen-presenting cells by chemerin, a novel processed ligand from human inflammatory fluids. J Exp Med. 2003;198:977-985.

9. Ernst MC, Sinal CJ. Chemerin: at the crossroads of inflammation and obesity. Trends Endocrinol Metab. 2010;21:660-667.

10. Wittamer V, Bondue B, Guillabert A, Vassart G, Parmentier M, Communi D. Neutrophil-mediated maturation of chemerin: a link between innate and adaptive immunity. J Immunol. 2005;175:487- 493.

11. Vermi W, Riboldi E, Wittamer V, et al. Role of ChemR23 in direct- ing the migration of myeloid and plasmacytoid dendritic cells to lymphoid organs and inflamed skin. J Exp Med. 2005;201:509- 515.

12. Parolini S, Santoro A, Marcenaro E, et al. The role of chemerin in the colocalization of NK and dendritic cell subsets into inflamed tissues. Blood. 2007;109:3625-3632.

13. Goralski KB, McCarthy TC, Hanniman EA, et al. Chemerin, a novel adipokine that regulates adipogenesis and adipocyte metabolism. J Biol Chem. 2007;282:28175-28188.

14. Bozaoglu K, Bolton K, McMillan J, et al. Chemerin is a novel adi- pokine associated with obesity and metabolic syndrome. Endocri- nology. 2007;148:4687-4694.

15. Rechache NS, Wang Y, Stevenson HS, et al. DNA methylation pro- filing identifies global methylation differences and markers of ad- renocortical tumors. J Clin Endocrinol Metab. 2012;97:E1004- E1013.

16. Velázquez-Fernández D, Laurell C, Geli J, et al. Expression profiling of adrenocortical neoplasms suggests a molecular signature of ma- lignancy. Surgery. 2005;138:1087-1094.

17. Fernandez-Ranvier GG, Weng J, Yeh RF, et al. Identification of biomarkers of adrenocortical carcinoma using genomewide gene expression profiling. Arch Surg. 2008;143:841-846; discussion 846.

18. Zheng Y, Luo S, Wang G, et al. Downregulation of tazarotene in- duced gene-2 (TIG2) in skin squamous cell carcinoma. Eur J Der- matol. 2008;18:638-641.

19. Lin W, Chen YL, Jiang L, Chen JK. Reduced expression of chemerin is associated with a poor prognosis and a lowed infiltration of both dendritic cells and natural killer cells in human hepatocellular car- cinoma. Clin Lab. 2011;57:879-885.

20. Pachynski RK, Zabel BA, Kohrt HE, et al. The chemoattractant chemerin suppresses melanoma by recruiting natural killer cell an- titumor defenses. J Exp Med. 2012;209:1427-1435.

21. Qu X, Han L, Wang S, et al. [Detection of chemerin and it’s clinical significance in peripheral blood of patients with lung cancer.]. Zhongguo Fei Ai Za Zhi. 2009;12:1174-1177.

22. Wang C, Wu WK, Liu X, et al. Increased serum chemerin level promotes cellular invasiveness in gastric cancer: a clinical and ex- perimental study. Peptides. 2014;51:131-138.

23. Zhang J, Jin HC, Zhu AK, Ying RC, Wei W, Zhang FJ. Prognostic significance of plasma chemerin levels in patients with gastric can- cer. Peptides. 2014;61C:7-11.

24. Giordano TJ, Kuick R, Else T, et al. Molecular classification and prognostication of adrenocortical tumors by transcriptome profil- ing. Clin Cancer Res. 2009;15:668-676.

25. de Reyniès A, Assié G, Rickman DS, et al. Gene expression profiling reveals a new classification of adrenocortical tumors and identifies molecular predictors of malignancy and survival. J Clin Oncol. 2009;27:1108-1115.

26. Demeure MJ, Coan KE, Grant CS, et al. PTTG1 overexpression in adrenocortical cancer is associated with poor survival and repre- sents a potential therapeutic target. Surgery. 2013;154:1405-1416; discussion 1416.

27. Chamberland JP, Berman RL, Aronis KN, Mantzoros CS. Chemerin is expressed mainly in pancreas and liver, is regulated by energy deprivation, and lacks day/night variation in humans. Eur J Endo- crinol. 2013;169:453-462.

28. Catalán V, Gómez-Ambrosi J, Rodríguez A, et al. Increased levels of chemerin and its receptor, chemokine-like receptor-1, in obesity are related to inflammation: tumor necrosis factor-& stimulates mRNA levels of chemerin in visceral adipocytes from obese patients. Surg Obes Relat Dis. 2013;9:306-314.

29. Herenius MM, Oliveira AS, Wijbrandts CA, Gerlag DM, Tak PP, Lebre MC. Anti-TNF therapy reduces serum levels of chemerin in rheumatoid arthritis: a new mechanism by which anti-TNF might reduce inflammation. PLoS One. 2013;8:e57802.

30. Gisondi P, Lora V, Bonauguri C, Russo A, Lippi G, Girolomoni G. Serum chemerin is increased in patients with chronic plaque psori- asis and normalizes following treatment with infliximab. Br J Der- matol. 2013;168:749-755.

31. Erdogan I, Deutschbein T, Jurowich C, et al. The role of surgery in the management of recurrent adrenocortical carcinoma. J Clin En- docrinol Metab. 2013;98:181-191.

32. Vinay DS, Ryan EP, Pawelec G, et al. Immune evasion in cancer: mechanistic basis and therapeutic strategies. Semin Cancer Biol. 2015;35(suppl):S185-S198.