Exogenous APN protects normal tissues from radiation-induced oxidative damage and fibrosis in mice and prostate cancer patients with higher levels of APN have less radiation-induced toxicities

Abstract

Radiation causes damage to normal tissues that leads to increased oxidative stress, inflammation, and fibrosis, highlighting the need for the selective radioprotection of healthy tissues without hindering radiotherapy effectiveness in cancer. This study shows that adiponectin, an adipokine secreted by adipocytes, protects normal tissues from radiation damage invitro and invivo. Specifically, adiponectin (APN) reduces chronic oxidative stress and fibrosis in irradiated mice. Importantly, APN also conferred no protection from radiation to prostate cancer cells. Adipose tissue is the primary source of circulating endogenous adiponectin. However, this study shows that adipose tissue is sensitive to radiation exposure exhibiting morphological changes and persistent oxidative damage. In addition, radiation results in a significant and chronic reduction in blood APN levels from adipose tissue in mice and human prostate cancer patients exposed to pelvic irradiation. APN levels negatively correlated with bowel toxicity and overall toxicities associated with radiotherapy in prostate cancer patients. Thus, protecting, or modulating APN signaling may improve outcomes for prostate cancer patients undergoing radiotherapy.

Graphical abstract

Highlights

• •Radiation reduces adiponectin plasma levels, which correlate with more radiation-induced toxicity in humans and mice.
• •Adiponectin treatment protects fibroblasts from damage and fibrotic progression in vitro.
• •Overexpression of adiponectin protects from chronic oxidative damage and fibrosis in vivo.
• •Adiponectin treatment does not confer protection to prostate cancer cells.

Abbreviations

(APN)

adiponectin

(ROS)

reactive oxygen species

(α-SMA)

alpha-smooth muscle actin

(AdipoR1 and AdipoR2)

adiponectin receptor 1 and 2

(AMPK)

adenosine monophosphate kinase

(PPAR)

peroxisome proliferator-activated receptor

(ACM)

adipocyte conditioned media

(HPrFs)

human primary prostate fibroblasts

(IBMX)

3-isobutyl-1-methylxanthine

(DMSO)

dimethyl sulfoxide

(MES)

2-[N-morpholino] ethanesulfonic acid

(SARRP)

small animal radiation research platform

(3-NT)

3-nitrotyrosine

(8-OHdG)

8-hydroxy-2′-deoxyguanosine

(4-HNE)

4-hydroxynonenal

(M.O.M.)

Mouse-On-Mouse

(H&E)

hematoxylin and eosin

(NGM)

fibroblast/normal growth media

(AdAPN)

adiponectin-expressing adenovirus

(AdEmpty)

empty adenoviral vector

(ULK1)

Unc-51 like autophagy activating kinase

(VPS34)

vacuole protein sorting 34

(Keap1)

Kelch-like ECH-associated protein 1

(Nrf2)

nuclear factor erythroid 2–related factor 2

(TGF-β)

transforming growth factor-beta

(TNFα)

tumor necrosis factor-alpha

(mTOR)

mammalian target of rapamycin

(interleukin-4)

IL-4

(ER)

endoplasmic reticulum

(PDI)

protein disulfide isomerase

(ERp44)

endoplasmic reticulum protein-44

(EBRT)

external beam radiation therapy

(EPIC-CP)

expanded prostate cancer index composite for clinical practice

(FACT-G)

functional assessment of cancer therapy-general

(FACIT-F)

functional assessment of chronic illness therapy fatigue subscale

1. Introduction

Radiotherapy is a vital component in the treatment strategies for a variety of cancer types/stages including, but not limited to, brain, head and neck, pelvic cancers (rectal, cervical and prostate), and breast cancer [[1], [2], [3], [4], [5], [6], [7], [8]]. Radiation causes acute damage to biomolecules resulting in double-stranded DNA breakage and water hydrolysis. Radiolysis of water results in the formation of highly reactive hydroxyl radicals, which generate a reactive oxygen species (ROS) cascade that causes secondary injury to DNA, proteins, and lipids [[9], [10], [11], [12]]. Radiation is an effective anti-cancer therapy because it allows for maximal dose targeting to the tumor and the impact of acute DNA damage is more detrimental to transcriptionally active, fast-dividing cells, like cancer [6,[13], [14], [15]].

Unfortunately, despite advancements in radiotherapy, such as, fractionation and utilizing non-static beam arrangements, radiation exposure still damages normal tissues, especially through the ROS cascade [16]. In addition to chronic oxidative damage to biomolecules in irradiated healthy tissue, radiation causes chronic inflammation and immune invasion, as well as activation of fibroblasts to myofibroblasts [[17], [18], [19], [20]]. Myofibroblasts have increased alpha-smooth muscle actin (α-SMA), collagen production, and extracellular matrix deposition, which contribute to chronic fibrosis [17,[21], [22], [23]]. Symptomatically, fibrosis manifests through the loss of proper organ function [17,20,24].

The rising incidence of cancer coupled with improving survival has led to an increase in the number of people living with treatment related complications [25]. This highlights the secondary complications of cancer treatment in recovering survivors and the need to consider the long-term effects of cancer treatments, like radiotherapy. Radiation protection is very important for prostate cancer patients for several reasons: the prostate is in close proximity to other pelvic organs and tissues (bladder, colon, rectum, periprostatic and other adipose deposits); it has a high survival rate as compared to other cancers, and chronic side effects of radiotherapy have been documented [17,20]. These chronic complications include incontinence, impotence, fibrosis, fatigue, and cancer recurrence [17,20,24,[26], [27], [28]]. This highlights the need for radioprotection that does not interfere with anti-cancer treatments and improves the future health and quality of life for cancer survivors.

Adiponectin (APN) is a 30 kDa protein that is secreted primarily by white adipose tissue, which serves as an endocrine and lipid storage organ [[29], [30], [31], [32], [33]]. Our lab has previously identified APN as a secreted protein with radioprotective effects on healthy fibroblasts [34]. APN's antioxidant, anti-fibrotic, and anti-inflammatory properties are well characterized [22,[35], [36], [37], [38], [39], [40], [41], [42], [43], [44]]. APN secretion is reduced in people with metabolic disorders, such as type 2 diabetes and obesity, who tend to have worse outcomes following cancer therapies [[45], [46], [47], [48], [49]]. APN is found in relatively high concentrations in human serum at ∼10–30 μg/mL in healthy people; however, in unhealthy individuals APN levels are significantly lower. APN signals through two primary receptors, APN receptors (AdipoR1 and AdipoR2). AdipoR1 activates adenosine monophosphate kinase (AMPK), a metabolic sensor that promotes catabolism, autophagy, and inhibits fibrosis [37,[50], [51], [52]]. AdipoR2 increases peroxisome proliferator-activated receptor alpha (PPAR-α) expression, a transcription factor associated with antioxidant defense [42,[53], [54], [55]]. Thus, APN plays a large role in signaling and maintaining normal metabolism in the human body.

In this current study, we demonstrate that APN levels inversely correlate with radiation toxicities in prostate cancer patients receiving radiotherapy and that APN levels are diminished during and after radiotherapy. Using our animal model of pelvic radiotherapy, we show that APN levels are reduced during and following radiation treatment. In addition, when APN is overexpressed in mice receiving pelvic radiation, there is less radiation-induced oxidative stress and fibrosis in the irradiated tissues. Thus, APN may play a key protective role during radiotherapy.

2. Materials and methods

2.1. Experimental animals

Wild type C57BL/6 mice (6–8 weeks of age) were obtained from Jackson Laboratories and kept on a 12h light and dark cycle. Animals were fed and watered, ad libitum, and housed at the University of Nebraska Medical Center (UNMC) in accordance with the Guide for Care and Use of Laboratory Animals by the National Institute of Health. Animal treatment procedures were approved and enforced by the UNMC Institutional Animal Care and Use Committee (20-019-03-FC).

2.2. Animal radiation treatment

For pelvic irradiation, C57BL/6 mice (6–8 weeks old) were anesthetized with a xylazine 11 mg/kg and ketamine 80 mg/kg solution intraperitoneally (ip). Their upper bodies were lead-shielded exposing the pelvis to x-irradiation at 2 Gy/min. Mice received 7.5 Gy for 5 consecutive days for a total dose of 37.5 Gy using a RadSource 2000 X-Ray Box Irradiator. The irradiated pelvic area was approximately 2 × 2 cm².

For radiation of implanted tumors, an image-guided Small Animal Radiation Research Platform (SARRP) was used to direct a 2 Gy x-ray dose (3 Gy/min) to the prostate tumors for 5 consecutive days (total dose of 10 Gy) using a 15 × 15 mm collimator.

2.3. Mouse serum collection

Approximately 50 μL of serum was collected from live mice via the saphenous vein. At harvest, blood was collected from the thoracic cavity immediately after euthanasia by severing the aorta. Collected blood was centrifuged at 3000×g for 10 min, and the serum was extracted and prepared for ELISA or immediately frozen and stored at −20 °C.

2.4. Enzyme linked immunosorbent assays (ELISAs) for mouse studies

Serum from C57BL/6 mice was evaluated for APN levels by ELISA. Mouse (Abcam, Ref. ab108785) and human (Abcam, Ref. ab99968) total APN kits were used according to the manufacturers' protocols to verify adenovirus APN overexpression. Mouse serum samples were diluted for each kit (1:1600 for mouse and 1:30,000 for human), and each assay was performed according to the manufacturers’ protocol. Standard curves were run with each set of samples to determine concentration. Absorbance was measured at 450 nm using a Tecan Plate Reader.

2.5. External damage scoring

Damage scoring was initiated 3 weeks after radiation treatment and was continued weekly. Animal scoring was blinded to treatment, and the same scorer was used throughout the experiment to maintain consistency. Mice were scored 0–1 point for hair loss, 0–1 point per side for leg-tucking (reduced mobility and tucked position), and 0–1 point per side for sores (0.5 dry, 1.0 moist desquamation) according to the skin irradiation damage scoring system [56].

2.6. Isolation and cell culture of mouse primary fibroblasts

Prostates from 6 to 8 week-old mice were isolated as described [34], placed in 5 mL Hank's Balanced Salt Solution (HBSS) (without Ca²⁺ and Mg²⁺) and quickly minced with a scalpel blade. The minced prostate tissue was then digested in 5 mg/mL type 1 collagenase in HBSS with Ca²⁺ and Mg²⁺ (Thermo Fisher, Ref. 17100014), until tissue fragments passed easily through a 2-mL pipette (≤30 min at 37 °C). The remaining tissue and cells were centrifuged at 500×g for 5 min, supernatant decanted, and cells were resuspended in Dulbecco's minimal essential media (DMEM), containing 10 % fetal bovine serum (FBS), 1 % penicillin/streptomycin, and 1 % non-essential amino acids. The primary prostate fibroblasts were cultured for no more than 2–3 weeks at 5 % CO₂ and 37 °C.

2.7. Isolation and cell culture of mouse primary adipocytes

Epididymal adipose tissues from 6 to 8 week-old mice were extracted as described [34], and digested in 2 mg/mL type 1 collagenase (Thermo Fisher, Ref. 17100014), agitating the tube every 5 min until the mixture became cloudy (≤45 min at 37 °C). The tissue and cells were centrifuged at 500×g for 5 min, separating the buoyant, mature adipose tissue and pelleting the stromal vascular fraction. The mature adipose tissue fraction was cultured in DMEM, containing 10 % FBS, 1 % penicillin/streptomycin, and 1 % non-essential amino acids allowing the mature adipocytes to migrate out of the tissue. The primary adipocytes were cultured for no more than 2–3 weeks at 5 % CO₂ and 37 °C.

2.8. Conditioned media

Media was changed every 2 days for the primary fibroblasts and adipocytes. Used media was collected and pooled during the culture. Before application or analysis, media was centrifuged to remove debris and cells at 700×g for 7 min. Media was stored at 4 °C and used within one week of collection.

2.9. Cell line culture

Human Primary Prostate Fibroblasts (HPrFs) were obtained from ScienCell (Ref. 4430). The 3T3-L1 (Ref. CL-173) cells and human prostate cancer cell-lines (PC-3, Ref. CRL-1435, LNCaP, Ref. CRL-1740, and C42B, Ref. CRL-3315) were obtained from American Type Culture Collection (ATCC). For HPrF culture, plates were pre-coated 24 h prior with 2 μg/cm² poly-l-lysine for 1 h at 37 °C. The cells were maintained under 80 % confluency, passaging, or changing media every 2 days in ScienCell Fibroblast Media (Ref. 2301) supplemented with 2 % FBS, 1 % Pen/Strep, and 1 % fibroblast growth supplement. PC-3, LNCaP, and C42B cells were cultured in HyClone RPMI-1640 medium supplemented with 10 % FBS and 1 % penicillin/streptomycin. Cells were regularly tested for mycoplasma contamination using the Lonza MycoAlert™ Mycoplasma Detection Kit (Ref. lT07-118).

2.10. Cell culture treatment

HPrFs or human prostate cancer cell lines (PC3 or C42B) were treated with AdipoRon, an APN receptor agonist, or dimethyl sulfoxide (DMSO), which is a vehicle control for AdipoRon. For cells treated with recombinant APN, phosphate buffered saline (PBS) was the vehicle control. AdipoRon was obtained from Sigma (Ref. SML0998). Recombinant human APN was purchased from BioRad (Ref. PHP196).

2.11. Cell culture irradiation

All irradiation of cultured cells was administered at 2 Gy/min using a Rad Source RS2000 X-ray box irradiator. The UNMC Radiation Safety Office performs dosimetry regularly. Copper filtration was used to attenuate the x-rays.

2.12. Quantitative RT-PCR

Primary mouse adipocytes or fibroblasts were pelleted by centrifugation (500×g, 5 min), mouse spleens were homogenized, and mRNA was isolated using the Qiagen RNeasy® Plus Mini Kit (Ref. 741340) according to the manufacture's protocol per sample type. The isolated mRNA was quantified and analyzed for purity by Tecan Infinite 200 Pro (Ref. 30050303) nano drop. Using the Power SYBR® Green RNA-to-CT™ 1-Step Kit (Applied Biosystems, Ref. 4389986), 40 ng of mRNA was combined with master mix to 20 μL total volume, and reactions were performed in a Bio-Rad C1000 Touch Thermocycler with the CFX96 Real-Time System as follows: 48 °C × 30 min, 95 °C for 10 min, 45 cycles of 95 °C for 15 s followed by 60 °C for 1 min, 65 °C for 5 s and 95 °C for 30 s. Mouse primers were obtained from Integrated DNA Technologies: AdipoR1 (F - 5′-TCT GCC TCA GTT TCT CCT GGC T-3′, R - 5′-GTA ATA GAG CCA CGG AAC GAA GC-3′), AdipoR2 (F - 5′-TCT TCC ACA CGG TGT ACT GCC A-3′, R - 5′-GGT AGA TGA AGC AAG GTT GTG GG-3′), T-bet (F-5′- TCC TGT CTC CAG CCG TTT CT-3’,R-5′- CGC TCA CTG CTC GGA ACT CT-3′), Gata-3 (F-5′- GGC GGC GAG ATG GTA CTG-3′, R-5′- TCT GCC CAT TCA TTT TAT GGT AGA -3′) and RPLP0 (F - 5′-GCA GGT GTT TGA CAA CGG CAG-3′, R - 5′-GAT GAT GGA GTG TGG CAC CGA-3′). C_T values were normalized to RPLP0 expression.

2.13. SDS-PAGE gel with western blotting

Lysates were made from harvested cells, and 40 μg of total protein was mixed with ddH₂O, Novex 4x dye and 10x sample reducing reagent and heated at 95 °C for 10 min. Samples were then run on an Invitrogen Bolt™ 4–12 % Plus (Ref. NW04120BOX) gel (200 V, 30 min) in 2-[N-morpholino] ethanesulfonic acid sodium dodecyl sulfate buffer. Proteins were transferred from the gels to nitrocellulose membranes via the semi-dry Invitrogen iBlot2 (Ref. IB21001) transfer apparatus. Even loading and complete transfer to membranes were evaluated with 2 min exposure to Ponceau followed by rinsing with water.

All membranes were then blocked for 1 h with either 5 % milk or 5 % bovine serum albumin (BSA) in Tris-buffered saline Tween (TBST) solution, according to the antibody manufacturers’ protocol. Primary antibodies were diluted 1:1000 in blocking solution and exposed to the membrane overnight at 4 °C. The membranes were washed 5 times in TBST for 5 min. Secondary antibody, diluted to 1:10,000 in blocking solution, was incubated on the membrane with shaking for 1 h at room temperature. Blots were then rinsed with TBST as before and developed with Thermo Pierce™ Enhanced ChemiLuminescence Western Blotting Substrate (Ref. 32106) to visualize relative protein levels. Quantification of band intensity was conducted using ImageJ software and normalized to Ponceau staining. Primary antibodies were acquired from Cell Signaling (AMPKα – 2532, phospho-AMPKα [T172] – 2535), Sigma (β-actin – A5441), and Novus Biotechnologies (PPARγ – NB600–636SS), and secondary antibodies were acquired from Invitrogen; anti-Rb conjugated to horseradish peroxidase (HRP, Ref. 31460) and anti-Ms HRP (Ref. A24524).

2.14. Population doublings

HPrFs were seeded at 45,000 cells/flask and allowed to proliferate for five days. Population doubling was then calculated (=3.32*(log(cell_d5)-log(cells_d0) by mixing diluted cells with HyClone™ Trypan Blue Solution (Ref. SV30084.01), which were counted on a hemocytometer by averaging the number of viable cells in the 4 corner fields. Cells were treated at day 0 with either sham or 5 Gy radiation and DMSO vehicle or 20 μM AdipoRon.

2.15. Clonogenic assays

Clonogenic assays on HPrF and prostate cancer cells (PC-3 and C42B) were performed as described [34]. Log phase HPrF cells were pre-treated, overnight, with either DMSO vehicle or 20 μM AdipoRon, and 1000–5000 (depending on treatment) cells were seeded in 6-well plates and irradiated with either 0 or 5 Gy. Log phase prostate cancer cells (PC-3 and C42B) were pre-treated, overnight, with either PBS vehicle or increasing doses of APN (0, 0.3, 3.0, and 6.0 ng/mL) and with 0 or 3 Gy x-irradiation, and 500-1000 cells were seeded in a 6-well plate. Colonies were allowed to form for 10 days, fixed, and counted with a dissecting microscope after staining with 5 % methanol and 0.5 % crystal violet solution. Plating efficiency and surviving fractions were then calculated.

2.16. Collagen contraction

HPrF cells in log phase growth were pretreated with PBS or 0.3 ng/mL of recombinant human APN 30 min prior to radiation with 0 or 3 Gy. On day 4 after treatment, cells were detached and mixed into a solution of rat tail collagen (Corning, Ref. 354249) diluted to 2 mg/mL in 0.5 M glacial acetic acid and 1X DMEM. A volume of 500 μL of the cell/collagen mixture per sample was then seeded to low-attachment 24-well plates. Once solid, 500 μL growth media was added to the wells, and a 10 μL pipette tip was used to detach the discs from the sides of the wells, allowing the disc to float in the growth media. The plate was then incubated at 37 °C for 6 h and disc area was measured using ImageJ.

2.17. yH2AX immunofluorescent staining

HPrF cells were pre-treated for 30 min with PBS vehicle or 0.3 ng/mL APN and then exposed to sham or 3 Gy irradiation. After an additional 30 min, cells were detached, and 8 x 10⁴ cells were spun on to slides and fixed in 4 % paraformaldehyde. Cells were permeabilized by in 0.5 % Triton X-100 in PBS for 8 min, washed 5 times in PBS for 5 min, and encircled with a hydrophobic marker on the slide. Blocking solution (10 % goat serum in PBS) was then applied to the cells and incubated for 1 h at 37 °C. Primary γH2AX antibody (Abcam, ab11174) was diluted 1:200 in blocking solution and applied to the cells for 1 h at 37 °C, followed by washing and application of the fluorescent secondary anti-rabbit antibody (AlexaFluor-488, Invitrogen a-11008) in blocking solution (1:500) for 1 h at 37 °C. After washing, cover slips were then applied with 20 μL 4′,6-diamidino-2-phenylindole (DAPI), and slides were imaged on a Leica DM 4000B LED fluorescent microscope at 20x after 30 min in the dark. A minimum of 6 fields per sample were imaged (selected from the DAPI channel only) and quantified to find a replicate mean. The experiment was performed 3 times and positive cells were identified with ImageJ by gating intensity and area of the nucleus stained.

2.18. Invivo administration of adenovirus

C57BL/6 mice were injected in the tail vein with either PBS (CON, RAD groups), PBS containing empty vector adenovirus at 5 × 10⁸ PFU/mouse (AdEmpty, RAD + AdEmpty), and PBS containing human APN adenovirus (AdAPN, RAD + AdAPN) as performed previously [35,57]. The AdAPN (Ref. 114145) and AdEmpty (Ref. 000047A) were acquired from Applied Biological Materials.

2.19. Tumor implantation

Athymic mice (Jackson Labs Nu/J Ref. 002019), 6–8 weeks old, were anesthetized with a combination of 2.5 % isoflurane and oxygen, administered continuously. After cleaning the surgical site with iodine, an incision was made in the middle of the lower abdomen to expose the bladder. The bladder was gently pulled externally until the prostate was visible. PC3 or LNCaP prostate cancer cells were injected into the dorsal prostatic lobe (50 μL in 50 % Matrigel containing 2.0 x 10⁶ cells/mouse). Incisions were closed with sutures and wound clips. Sterile surgical procedures were maintained at all times under a laminar flow hood. Mice were given 0.1 mg/kg buprenorphine SR for pain subcutaneously. Mice were monitored until they regained consciousness and then daily until removal of the wound clips. Tumors were allowed to grow to 200 mm³ (measured with the SARRP cone-beam computed tomography function) [58] and were treated with AdEmpty or AdAPN (5 × 10⁸ PFU/mouse, ip) 3 days prior to irradiation. Tumors were measured weekly and at harvest.

2.20. Tissue harvest

Two months post pelvic irradiation, spleens, abdominal skin, bladders, left and right epididymal fat pads, prostates, and serum were harvested. Spleens were processed for mRNA extraction and blood for ELISA. The remaining tissues were fixed in 4 % formalin. After fixation, sections were stored in 70 % ethanol and then embedded in paraffin. Tissue sections were cut for further analysis and staining by the UNMC Core Tissue Facility.

2.21. Tissue immunohistochemistry

Fixed tissue sections of the skin, bladder, and fat were deparaffinized and stained for markers of ROS/RNS with 3-nitrotyrosine (3-NT) (Abcam, Ref. ab61392, 1:1000) for protein damage, 8-hydroxy-2′-deoxyguanosine (8-OHdG) (Abcam, Ref. ab62623, 1:200) for DNA damage, and 4-hydroxynonenal (4-HNE) (R&D Systems, Ref. MAB3249, diluted to 0.5 μg/mL) for lipid peroxidation. A Mouse on Mouse (M.O.M.) kit from Vector Labs (Ref. BMK-2202) was utilized according to the manufacturer's instructions to minimize non-specific staining. Appropriate fluorescent secondary antibodies from Invitrogen diluted 1:500, AlexaFluor 488 goat anti-mouse (Ref. A-28175, green) or AlexaFluor 555 goat anti-mouse (Ref. A-28180, red) were added. After washing, coverslips were mounted with ProLong Gold Antifade with DAPI and imaged using a Leica DM 4000B LED fluorescent microscope. Further analysis of intensity and area was conducted using ImageJ software.

2.22. Trichrome staining

Deparaffinized mouse skin and bladder sections were refixed in Bouin's solution at room temperature for 16 h. Sections were stained with Weigert's iron hematoxylin for 10 min for nuclei and then washed in warm running water for 10 min, 1 % Biebrich Scarlet-Acid Fuchsin for 5 min for cytoplasm, keratin, and muscle, then differentiated in 2.5 % phosphomolybdic- 2.5 % phosphotungstic acid solution for 15 min. Finally, 2.5 % Aniline blue was added for 8 min to visualize collagen, followed by differentiation for 1 min in 1 % glacial acetic acid. Sections were dehydrated through graded alcohols and cleared with xylenes, then mounted with an aqueous medium. Sections were imaged using a Leica DM 4000B microscope. A minimum of six fields were collected per animal and averaged. ImageJ was used to quantify percent collagen in the bladder and skin connective tissue. Epidermal thickness was likewise measured, assessing the width of the red epidermal layer at 25 equally spaced points in each field, excluding hair follicles.

2.23. Adipocyte area

Sections of adipose tissue were hematoxylin and eosin (H&E) stained to visualize adipose tissue morphology. Adipocyte size was measured using ImageJ to trace the cells and calculate the average area of the cells in 6 fields/mouse.

2.24. Prostate cancer trial

Data for this study were obtained from an ongoing descriptive, longitudinal IRB-approved (# 0504-18-EP) study investigating men with non-metastatic prostate cancer receiving external beam radiation therapy (EBRT). Written informed consent was provided by all participants. Men were included if they: (a) had a diagnosis of non-metastatic prostate cancer, (b) were receiving androgen-deprivation therapy, (c) were scheduled to receive EBRT, and (d) were at least 19 years of age.

Data were collected at 6 study visits over 2 years: baseline (A), midpoint of EBRT (B), completion of EBRT (C), and 6 months (D), 1 year (E) and 2 years (F) post completion of EBRT. Data were obtained from participant medical records, self-report questionnaires (cancer-related fatigue and health-related quality of life), and blood samples.

Medical Record Data. Medical records were used for assessment of demographic (age, race, ethnicity) and disease-related characteristics (Gleason score and Grade Group, prostate cancer risk group, radiation dose and fractions, and ADT treatment length).

Fatigue. The Functional Assessment of Chronic Illness Therapy Fatigue subscale (FACIT-F) is a 13-item self-reported questionnaire that is validated in the oncology population. FACIT-F has good test-retest reliability (r = 0.90), internal consistency reliability (α = 0.93 and 0.95) on initial and test-retest administration, suggesting that it can be administered as an independent, unidimensional measure of fatigue. Higher scores indicate less fatigue [59]. To capture fatigue triggered by EBRT, participants were phenotyped (fatigued versus non-fatigued) using common phenotyping approaches [60].

Health-Related Quality of Life. Health-related quality of life was evaluated using two self-report questionnaires, the Expanded Prostate Cancer Index Composite for Clinical Practice (EPIC-CP) and the Functional Assessment of Cancer Therapy- General (FACT-G). The EPIC-CP is a 16-item self-reported questionnaire designed to evaluate health-related quality of life in prostate cancer patients [61,62]. The EPIC-CP provides an overall prostate cancer quality of life score as well as 5 symptom domain (urinary incontinence, urinary irritation/obstruction, bowel, sexual, and vitality/hormonal) scores. The EPIC-CP has good internal consistency (α = 0.64–0.84) [61]. Higher scores on the EPIC-CP equate to worse health-related quality of life [61]. The FACT-G (version 4) is a well-validated 27-item self-reported questionnaire that assesses health-related quality of life across 4 different well-being domains: physical (7 items), social (7 items), emotional (6 items), and functional (7 items). Higher scores indicate better quality of life [63,64].

2.25. Human serum collection

Blood samples were collected at each time point via a peripheral blood draw. Two 4-mL BD Vacutainer® CPT™ Mononuclear Cell Preparation Tube - Sodium Citrate was used to collect peripheral blood for buffy coat and serum extraction and stored at −80 °C.

2.26. APN ELISA for human samples

For detection in human plasma samples, a human APN ELISA (R&D Systems, Ref. SRP300) was used according to manufacturer's instructions. Plasma samples were diluted 1:100 and measured at 450 nm using a Tecan Plate Reader with a standard curve to determine APN levels in plasma.

2.27. Statistics

For human clinical specimens, generalized estimating equation (GEE) approach was used to manage correlation between measurements at different times. The robust variance estimation was used to manage potential non-normality and variance heterogeneity when computing Wald test-based p-values for all regression models. All significance levels were set at 0.05. Regression models of APN level on ADT status adjusting for measurement time were used to determine whether the APN decreased when patients stopped ADT. Regression modeling of APN levels and baseline APN levels on baseline BMI were used to study their association. Since ADT increases serum APN levels [65], we excluded any data collected after cessation of ADT or when ADT status was not known for the following regression analysis. Regression modeling of APN levels at specific time points were used to determine whether there was a decrease of APN during and after radiation therapy with time points pooled together or evaluated separately. Separate regression models of symptom, quality of life scores including their sub scores as well as fatigue score on baseline APN, total APN, as well as their dichotomized version at the cut point of 7500 ng/mL were run with the adjustment of measurement time to evaluate the association between APN and these scores. Regression of APN on dichotomized fatigue score at the cut point level of 34, time, and their interaction are performed to test whether there were differences in APN as compared to fatigued and non-fatigued patients over time.

For the rest of the experimental data, statistical analysis for all experiments were conducted in Prism GraphPad v.8 software. Graphical representations show the data mean and error bars to represent the standard deviation. Statistical significance was determined by student's t-test when differentiating between only 2 groups. For experiments with greater than 2 groups, a 1-way ANOVA (alpha = 0.05) was conducted followed by a Tukey's test to determine the p values for multiple comparisons.

3. Results

3.1. Prostate fibroblasts express receptors, AdipoR1 and R2, and activate APN signaling

Given that radiation induces many chronic side effects in prostate cancer patients through fibrosis, we wanted to determine the effects of APN on fibroblasts. We verified that prostate fibroblasts express AdipoR1 and AdipoR2 mRNA, the two primary APN receptors (Fig. 1A). We investigated the effects at 24 h post-radiation and did not observe significant effects on AdipoR1 or AdipoR2 expression (data not shown). As AdipoR1 and R2 are reported to show differential signal transduction, we measured phosphorylation of AMPK for its association with AdipoR1 signaling [54] (Fig. 1B). Prostate fibroblasts exhibited a significant increase in AMPK phosphorylation after 12h treatment with either recombinant APN or adipocyte conditioned media as compared to the treatment with fibroblast media (NGM). Phosphorylated AMPK was normalized to total AMPK levels. Ponceau was used to normalize protein loading. PPAR-α has been reported to be increased upon AdipoR2 activation [55]. We observed elevated PPAR-α levels in our APN-treated samples compared to vehicle control at 24 h, using β-actin as a loading control (Fig. 1C). This data supports that fibroblasts respond to receptor-mediated APN signaling.

Fig. 1.

Prostate fibroblasts express AdipoR1 and AdiopR2 and APN treatment protects human prostate fibroblasts from radiation damage. (A) RT-PCR was used to measure the presence and relative abundance of AdipoR1 and AdipoR2 mRNA in primary mouse prostate fibroblasts. (B) Representative blots with loading control (Ponceau). Phosphorylation of AMPK (normalized to Ponceau and total AMPK) was measured by Western blot in prostate fibroblasts after 12 h treatment with either adipocyte conditioned media (ACM), or 0.3 ng/mL APN (APN). (C) Representative PPARα blots with loading control (β-actin). Fibroblasts were treated for 24 h with normal growth media (NGM) and PBS vehicle or 0.3 ng/mL APN (APN). Average band densitometry normalized to loading controls is shown below each band. (D) Primary human prostate fibroblast (HPrF) proliferation was measured by counting the number of population doublings over 5 days following either 0 Gy sham- or 5 Gy irradiation with PBS vehicle or the APN receptor agonist, AdipoRon (20 μM). (E) A clonogenic assay was performed on HPrF cells treated with radiation and vehicle or radiation and AdipoRon prior to plating and surviving fractions were measured. (F) A collagen contraction assay was performed at 0 and 3 Gy irradiated groups that had either been treated with 0.3 ng/mL APN or PBS-vehicle. The treated cells were mixed into the collagen disc, and disc size was measured 6 h later. (G) HPrF cells were pre-treated with 0.3 ng/mL APN or PBS-vehicle 30 min prior to irradiation. Cells were administered 0 or 3 Gy of radiation and fixed and stained for γH2AX (green). A minimum of six fields/biological replicate (n = 4) were averaged for quantification. White bar represents 100 μm. Panels A and C were analyzed with a t-test, while panel B and D-G were analyzed with a 1-way ANOVA followed by Tukey's multiple comparison test for significance. * indicates a significant difference (p < 0.05, n = 3, unless otherwise indicated) from the control. # represents a significant difference (p < 0.05, n = 3, unless otherwise indicated) from the radiation group. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.2. APN protects human prostate fibroblasts from damage and activation caused by radiation

Our lab previously reported that primary mouse adipocyte conditioned media and recombinant APN treatment protected primary mouse fibroblasts from the effects of radiation in vitro [34]. We next tested similar endpoints in primary fibroblasts from human prostate tissue (HPrFs). We examined the HPrF proliferation by measuring population doubling over 5 days (Fig. 1D). To induce cell cycle arrest, cells were exposed to 5 Gy radiation, which resulted in a significant loss in proliferation. Proliferation was significantly restored with pretreatment of the AdipoR1 and AdipoR2 agonist, AdipoRon (20 μM) for 1 h. AdipoRon was used to determine if activating the AdiopR1 and R2 receptors was sufficient to protect fibroblasts through radiation exposure. We also measured the growth of HPrFs using a clonogenic assay, in which the 5 Gy-treated fibroblasts had significantly lower surviving fractions that were partially recovered with AdipoRon treatment (Fig. 1E).

APN signaling has been linked to reduced fibrotic markers in skin fibroblasts, such as, collagens and α-SMA [17,66,67]. To test for myofibroblast differentiation, a 3 Gy dose of radiation was used to treat HPrFs to induce differentiation rather than cell death. The cells were then seeded in collagen to evaluate the ability of the treated fibroblasts to contract the discs (Fig. 1F). Higher collagen contractility indicates myofibroblast differentiation, which is a pro-fibrotic phenotype. Irradiated fibroblasts demonstrated increased contractility, indicating myofibroblast properties. APN treatment of irradiated prostate fibroblasts significantly prevented fibroblast contractility and was indistinguishable from unirradiated control samples.

Another effect of radiation is DNA damage. To observe the effects of APN treatment on DNA damage, irradiated HPrFs were treated with PBS vehicle or 0.3 ng/mL APN 30 min prior to irradiation (3 Gy). Thirty minutes post-radiation, cells were stained for γH2AX, a DNA repair protein that indicates double-stranded DNA breaks (Fig. 1G). While radiation resulted in a four-fold increase in γH2AX-staining, this was reduced by approximately 50 % with APN treatment.

3.3. Mice overexpressing APN exhibit reduced signs of external radiation damage

To determine if these radioprotective effects also occur in vivo, we overexpressed circulating APN by IV injection of human APN-expressing adenovirus (AdAPN) or empty vector (AdEmpty) at 5 × 10⁸ PFU/animal in C57BL/6 mice [35,57] according to Fig. 2A. The normal human range of circulating APN is approximately 10–30 μg/mL, though the range decreases with unhealthy adipose tissue [68,69]. APN levels measured lower in C57BL/6 mice, ranging from ∼5 to 15 μg/mL (Fig. 2B). The dosing scheme was developed from an AdAPN pilot study, in which maximum APN overexpression was measured at 3 days post-injection and lasted approximately two weeks using ELISAs for human and mouse APN (Fig. 2B). At baseline, plasma APN concentration in C57BL/6 mice was ∼8 μg/mL of serum APN. At 3 days post-injection, AdEmpty mice showed no change from baseline and the AdAPN-treated mice had APN levels close to 20 μg/mL.

Fig. 2.

Mice overexpressing APN have higher levels of circulating APN, reduced external damage from radiation, and present with less Th2-skewing in splenocytes following radiation. (A) An experimental timeline is shown, detailing the pelvic irradiation (RAD) (7.5 Gy × 5 days) and empty (AdEmpty) or human APN (AdAPN) adenovirus (5 × 10⁸ PFU/mouse, IV) treatments. (B) Total APN levels in the serum of control, AdEmpty, and AdAPN-treated mice 3 days post-injection. (C) External radiation damage to the animals (n = 10/group) was measured weekly, starting 3 weeks following the first radiation dose. Insert pictures are representative images of mice treated with AdAPN or AdEmpty 3 weeks after radiation exposure. External damage was measured with 5 representing the highest indicators of damage and 0 being no observable damage. Mice were scored according to established radiation scoring parameters (hair loss, sores, erythema, moist desquamation) [59]. * indicates a significant difference (p < 0.05) from the control. To visualize phenotypic skewing of T-helper cells in the spleens of treated and untreated mice, (D) Gata-3 (Th2) mRNA levels were quantified. (E) T-bet (Th1) mRNA levels were measured. Spleens were collected from all (n = 5, CON and n = 4, RAD) (n = 10 for other groups). * indicates a significant difference (p < 0.05) from the control.

Mice were injected with AdAPN or AdEmpty 3 days prior to radiation. Mice received 0 Gy × 5 days sham irradiation for controls or 7.5 Gy × 5 days of x-ray irradiation to the pelvic region using shielding and a RS2000 box irradiator. After 3 weeks, the animal skin was scored weekly according to the guidelines specified in Radiobiology for the Radiologist (6th Edition) from zero to five according to hair loss, sores, erythema, and moist desquamation [56] (Fig. 2C). A score of 5 indicates maximal bilateral damage to the irradiated area, while a 0 score indicates no observable damage. Radiation resulted in skin damage, while the APN treated animals had significantly less skin damage throughout the experiment.

3.4. APN overexpression reduces mRNA indicators of Th2 skewing of T-helper cells in mouse spleens following radiation

Both APN and radiation are reported to modulate the immune system. APN reduces inflammation, while radiation promotes a Th2 wound healing response or immune environment associated with tumor development [[70], [71], [72], [73]]. Furthermore, inflammation is implicated in fibrosis, which is a phenotypic endpoint of radiation damage to normal tissue [[74], [75], [76]]. To assess for the effects of radiation and APN on immune response, mRNA was harvested from mouse spleens. Levels of Gata-3 mRNA, an indicator of Th2 type T-cells, were elevated in the radiation and RAD + AdEmpty groups but not in the RAD + AdAPN animals (Fig. 2D). Radiation and/or APN treatment did not significantly alter mRNA levels of the Th1 indicator, T-bet (Fig. 2E). Thus, radiation caused a significant Th2:Th1 immune shift that was not observed in the RAD + APN treatment group. These data suggest that APN overexpression inhibits the Th2 skewing caused by radiation exposure.

3.5. Markers of oxidative stress caused by radiation exposure to normal tissues are reduced by APN treatment

Given the trend of reduced external damage and reduced Th2 skewing due to APN treatment, pelvic tissues from these mice were further analyzed. Radiation causes chronic increases in ROS, primarily due to enhanced NADPH oxidase activity and increased mitochondrial ROS generation [17,18,23,77]. Sustained increases in ROS levels can induce fibrosis and DNA damage independent of direct radiation damage [78,79]. Previously, we have shown at 2 months post-irradiation, both skin and bladder have increased collagen, a hallmark of fibrosis [23]. To investigate whether APN overexpression could prevent chronic oxidative and nitrosative stress contributing to fibrosis, skin was collected from the abdomen of each animal and stained for 3-NT, an indicator of nitrosative and oxidative damage to proteins (Fig. 3A). Radiation alone or with AdEmpty resulted in a ∼2-fold increase in 3-NT staining, but the 3-NT levels in the RAD + AdAPN animals were indistinguishable from unirradiated controls (Fig. 3B). Bladder sections were stained for 8-OHdG, a marker for oxidative DNA damage (Fig. 3C). A ∼2-fold increase in 8-OHdG staining was observed in mouse bladders from the RAD and RAD + AdEmpty groups. In the RAD + AdAPN group, 8-OHdG staining was indistinguishable from the 0 Gy control group (Fig. 3D). These results demonstrate the chronic nature of oxidative damage caused by radiation exposure to normal tissues, and the radioprotective potential of APN overexpression.

Fig. 3.

Markers of radiation-induced oxidative stress are lower in the skin and bladders of animals overexpressing APN. Skin sections from control and irradiated mice were collected and stained for 3-NT, an indicator of nitrosative/oxidative protein damage (3-NT = green, DAPI = blue). (A) Representative images for the CON, RAD, RAD + AdEmpty, and RAD + AdAPN are shown and (B) quantified. (C) Bladder sections were stained for 8-OHdG to indicate chronic oxidative DNA damage (8-OHdG = green, DAPI = blue) and (D) quantified. A minimum of six fields/sample were collected randomly and analyzed for positive staining area (>6 fields/mouse, 10 mice/group). White bar represents 100 μm * indicates a significant difference (p < 0.05) from the control. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.6. Fibrosis of healthy pelvic tissues following radiation exposure is reduced with APN overexpression

To examine signs of fibrosis and damage in normal pelvic tissue, bladder sections were trichrome stained and analyzed to estimate collagen deposition (Fig. 4A). Radiation alone or with AdEmpty caused a significant increase in the percent collagen measured in mouse bladder muscle 2 months after pelvic irradiation. Irradiated animals overexpressing APN, however, had collagen deposition equivalent to the unirradiated animals (Fig. 4B). Abdominal skin sections were trichrome stained (Fig. 4C) and evaluated for epidermal thickness (Fig. 4D) and percent collagen (Fig. 4E). Epidermal thickening from radiation was significantly reduced with APN treatment. The increase in collagen density in the skin caused by radiation was not seen in samples from irradiated animals treated with APN. Trichrome tissue analysis demonstrates that signs of fibrosis in the skin and bladders of irradiated mice are significantly reduced by overexpressing APN.

Fig. 4.

Irradiated bladder and skin tissues show signs of fibrosis, except when treated with APN. Bladder and skin tissues from treated and untreated mice were collected and trichrome stained (blue = collagen, red = epithelium/muscle). (A) Representative bladder images are shown for 0 Gy control, RAD + AdEmpty, and RAD AdAPN, and (B) the percent collagen was quantified for all groups. Arrows indicate areas of enhanced collagen deposition. (C) Representative images of skin sections are shown of 0 Gy control, RAD + AdEmpty, and RAD AdAPN, and (D) epidermal thickness and (E) the percent collagen in the dermis was quantified for all groups. The percent collagen in the bladder and skin in the RAD + AdAPN group was indistinguishable from the CON group. Black bar indicates 100 μm * indicates a significant difference (p < 0.05, n = 10) from the control. # represents a significant difference (p < 0.05, n = 10) from the radiation group. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.7. APN overexpression protected mice from radiation damage of epididymal white adipose tissue following pelvic irradiation

As mature white adipose tissue is the primary producer of endogenous APN, RT-PCR was used to verify the expression and relative abundance of AdipoR1 and AdipoR2 in primary adipocytes (Fig. 5A). mRNA for both receptors were present at similar levels, indicating that adipocytes can induce APN receptor signaling in an autocrine manner. Therefore, we harvested the epididymal fat pads from the pelvic region of different groups of mice. Sections of adipose tissue were H&E stained to visualize tissue morphology (Fig. 5B). The fat from the RAD and RAD + AdEmpty animals had significantly smaller adipocytes than the unirradiated groups (Fig. 5C). Although adipocytes from the RAD + AdAPN animals were significantly smaller than the control group, the fat cell area was still approximately 3 times larger than the RAD and RAD + AdEmpty samples. Sections were also stained for markers of oxidative stress, 8-OHdG for oxidative DNA damage (Fig. 5D) and 4-hydroxynonenal (4-HNE) for lipid peroxidation (Fig. 5F). Radiation caused a significant increase in both markers of oxidative damage to DNA (Fig. 5E) and lipids (Fig. 5G). AdAPN-treated mice had significantly less DNA and lipid oxidation in the adipose tissues as compared to mice receiving only radiation (Fig. 5E and G). This data shows that APN overexpression can protect adipose tissue from signs of radiation damage, but also, that adipose tissue, a major source of endogenous APN, is particularly sensitive to damage from radiation.

Fig. 5.

Mouse epididymal white adipose tissue deposits are damaged by radiation but protected with APN. (A) Relative abundance of AdipoR1 and AdipoR2 mRNA detected in primary mouse adipocytes by RT-PCR (n = 3). (B) Representative images of mouse adipose tissue (Control, RAD, RAD + AdEmpty and RAD + AdAPN), and (C) fat cell area is quantified for all groups. White adipose tissue sections were also stained for markers of (D) oxidative damage to DNA (8-OHdG = red, DAPI = blue) and (F) lipid peroxidation (4-HNE = red, DAPI = blue). Images for the control, RAD, RAD + APN and the IgG staining-control are shown. (E) Percentage of cells staining positive for 8-OHdG and (G) mean fluorescent intensity (MFI) for 4-HNE is quantified. Black bar indicates 50 μm. White bar represents 100 μm * indicates a significant difference (p < 0.05, n = 10) from the control. # represents a significant difference (p < 0.05, n = 10) from the radiation group. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3.8. APN treatment does not protect prostate cancer cells from radiation damage

The effects of overexpressing APN need to be evaluated in prostate cancer models when combined with radiation. Dose-escalating clonogenic assays in vitro were conducted on the aggressive, androgen-independent PC-3 (Fig. 6A) and C42B cells (Fig. 6B). For both cell lines, increases in APN dose did not improve the surviving fractions of the unirradiated or irradiated cancer cells. In fact, a small but significant decrease in growth was observed in both unirradiated and irradiated prostate cancer cells, which agree with our previous findings [34]. To investigate the effects of radiation and APN treatment on tumors in vivo, PC-3 (Fig. 6C) and the less aggressive, androgen-dependent LNCaP cells (Fig. 6D) were orthotopically implanted in the prostates of nude mice. At 6.5 weeks post-implantation, AdAPN or AdEmpty mice were injected 3 days prior to radiotherapy (2 Gy for 5 days). Again, for both cell lines, AdAPN treatment did not confer a growth advantage to the implanted prostate tumors. Although not significant, there was a trend of reduced tumor growth with AdAPN treatment.

Fig. 6.

APN treatment does not protect prostate cancer from radiotherapy. Clonogenic assays were performed on (A) PC-3 and (B) C42B prostate cancer cells. The cells received either 0 or 3 Gy and were then treated with increasing doses of recombinant human APN (0, 0.3, 3.0, and 6.0 ng/mL, n = 3). Human prostate tumor cells, (C) PC-3 and (D) LNCaP, were orthotopically implanted into the prostates of nude mice and allowed to grow. Once tumors reached ∼200 mm³, mice were then treated 2 Gy irradiation for 5 days and treated with either AdEmpty or AdAPN (5 × 10⁸ PFU/mouse). Tumor volumes were measured invivo until the final measurement at harvest, which was measured exvivo (n = 10).

We also investigated oxidative damage markers in tumor and surrounding normal prostate tissues with and without APN treatment. We found that APN treatment had little effect on 4-HNE or 3-NT staining in the tumor (Supplemental Fig. 1&2). However, in surrounding normal prostate tissues, 3-NT was significantly reduced with APN treatment and there was a trend of reduced 4-HNE adducts with the addition of APN (Supplemental Fig. 1&2). One reason why APN may have selective effects on tumor vs normal tissues is that prostate tumors have reduced APN receptors [80]. Another reason could be that APN inhibits the mammalian target of rapamycin (mTOR) signaling pathway, which is upregulated in prostate cancer cells but not in normal prostate cells [81]. The mTOR signaling pathway is crucial to promote cancer survival, growth, and aggressiveness. mTOR levels were significantly reduced from 15 min to 24 h post treatment with APN in PC-3 cancer cells (Supplemental Fig. 3).

3.9. APN secretion is altered by radiation exposure and reduced levels of APN correlate to more radiation-induced skin toxicity in mice

Given the changes observed in adipose tissues when exposed to radiation, we tested the effects of pelvic irradiation on mouse serum APN levels. We measured the APN levels in the serum of the animals at varying intervals following radiation treatment (0 Gy and 7.5 Gy × 5 days, Fig. 7A). At 2 months post-irradiation, serum APN levels in the RAD group were significantly reduced by ∼50 % from the control group. This significant reduction in circulating APN was also seen at 6 months, indicating that there is a chronic reduction of APN serum levels following radiation. These results agree with a small clinical study of childhood cancer survivors in which radiation resulted in a lower trend of serum APN years after radiation therapy [70]. We then correlated external damage scores of irradiated skin after 2 months with APN levels and found a significant negative correlation; high APN levels corresponded with lower damage scores, while low APN levels trended toward higher external damage (Fig. 7B).

Fig. 7.

Adipocyte secretion of APN is altered by radiation exposure. (A) Time course of serum APN levels in irradiated C57Bl/6 mice out to 6 months (n = 6). (B) External damage scores from individual mice were compared with serum APN levels (n = 10). Regression analysis performed with Prism GraphPad (R² = 0.8, p = 0.0011).

3.10. APN levels decrease during treatment in prostate cancer patients undergoing external beam radiation therapy (EBRT)

We then wanted to determine whether radiation causes a reduction of APN levels during and after radiotherapy in prostate cancer patients and if APN levels correlated with toxicities associated with radiation exposure. A non-therapeutic clinical trial was performed on prostate cancer patients receiving ADT and EBRT to better understand if a correlation exists between plasma APN levels and clinical outcomes from standard of care. Prostate cancer patient demographics from this study are found in Supplemental Table 1. Plasma samples were taken from prostate cancer patients before radiation, midway through radiation, at the completion of radiotherapy, and approximately 6 months, 1 year, and 2 years post-radiation. At each of these timepoints, patients also completed EPIC-CP, FACT-G, and FACIT-F questionnaires. All patients had started androgen deprivation therapy (ADT) prior to radiotherapy. Since ADT is known to increase plasma APN levels [65] and cause symptoms that may overlap with those induced by radiation, we excluded any data collected after cessation of ADT or when ADT status was not known. Indeed, we found that APN decreased an average of 1921 ng/mL (p = 0.048) when patients stopped ADT. Furthermore, it is well known that APN levels negatively correlate with body mass index (BMI). Accordingly, in our cohort baseline APN levels (p ≤ 0.001) were negatively correlated with BMI (Supplemental Fig. 4). Therefore, the data obtained from our cohort is congruent with previous studies conducted on plasma APN.

To investigate whether radiation influences plasma APN levels in pelvically irradiated humans, a change in APN from baseline was calculated. We found that APN levels were significantly reduced by an average of 556 ng/mL (p = 0.009) during and after radiotherapy. Specifically, significant decreases were observed at the midpoint (p = 0.029) and completion (p = 0.027) of radiation therapy.

3.11. Patient-reported symptom and quality of life scores are correlated with APN levels at, prior to, and throughout radiotherapy

We then investigated the relationship between APN levels and questionnaire scores. We found that APN levels were inversely correlated with urinary irritation/obstruction (p = 0.021, Fig. 8A), bowel (p = 0.002, Fig. 8B), hormonal (p = 0.024, Fig. 8C), and overall (p = 0.005) symptom scores (Fig. 8D), where a higher score is indicative of more symptoms. We observed no relationship between APN and sexual symptoms (Supplemental Table 2).

Fig. 8.

Plasma APN levels negatively correlate with radiation toxicities and fatigue in prostate cancer patients at baseline, during radiotherapy and after radiotherapy. (A) Plasma APN levels negatively correlate with urinary irritation and obstruction. Insert shows urinary irritation/obstruction symptoms in patients with APN >7500 ng/mg vs. patients with APN<7500 ng/mL at three time points. (B) Plasma APN levels negatively correlate with bowel symptom scores. Insert shows bowel symptoms in patients with APN >7500 ng/mg vs. patients with APN<7500 ng/mL at three time points. (C) Plasma APN levels negatively correlate with vitality/hormonal symptom scores. Insert shows vitality/hormonal symptoms in patients with APN >7500 ng/mg vs. patients with APN<7500 ng/mL at three time points. (D) Plasma APN levels negatively correlate with overall symptom scores. Insert shows overall symptoms in patients with APN >7500 ng/mg vs. patients with APN<7500 ng/mL at three time points. (E) Patients that are fatigued (FACIT-F score <34) had significantly lower APN prior to and during radiation as compared to non-fatigued patients (FACIT-F score >34). A = prior to radiotherapy, B = midpoint in radiotherapy, C = completion of radiotherapy, D = 6 months post-radiotherapy, E = 1 year post-radiotherapy, and F = 2 years post-radiotherapy. The (*) symbol denotes p ≤ 0.05 at specified time points.

Based on clustering in our data and reported reference ranges [82,83], we dichotomized patients into high and low APN groups using a cutoff value of 7500 ng/mL and found that the low APN group had significantly higher symptom scores. Specifically, low APN was associated with worse urinary obstruction/irritation (p = 0.011, Fig. 8A insert), bowel symptoms (p = 0.001, Fig. 8B insert), vitality/hormonal symptoms (p = 0.015, Fig. 8C insert), and increased overall symptoms (p = 0.007, Fig. 8D insert). Additionally, the low APN group at baseline had significantly lower symptom-related (physical) and overall quality of life (p = 0.03 and 0.015, respectively, Supplemental Table 3).

Finally, using a previously validated definition of fatigue (FACIT-F score <34), patients were dichotomized into fatigued and non-fatigued groups. We found that fatigued patients had significantly lower APN levels than non-fatigued patients at baseline (p = 0.001) and at the midpoint of radiotherapy (p = 0.005), but not at completion of radiation (Fig. 1E). Furthermore, APN was positively associated with reduced fatigue at baseline and throughout radiotherapy (p = 0.014 and 0.008 respectively, Supplemental Table 2).

Overall, these data demonstrate that lower plasma levels of APN correlate with increased radiation toxicities. They also show that patients undergoing pelvic irradiation have reduced APN during and following radiation exposure.

4. Discussion

As treatment of cancer has become more successful, it is imperative that our treatment methods should extend beyond treating the primary tumor to addressing chronic complications associated with treatment. Radiation exposure to normal tissue causes damage in the form of DNA double-stranded breaks and hydrolysis of water resulting in a persistent ROS cascade [9,10,12]. Normal tissues exposed to radiation are chronically oxidatively and nitrosatively damaged, inflamed, and fibrotic. In patients, irradiated pelvic tissues that have become fibrotic manifest in bladder and bowel incontinence, erectile dysfunction, and bowel fibrosis [17,19,77]. Thus, there is a need to protect normal tissues during radiotherapy and to identify individuals at increased risk of developing radiotherapy-induced toxicities.

In this current study, we showed that APN levels are significantly reduced acutely and chronically after pelvic irradiation in both mice and humans. Specifically, low APN concentrations in the serum were associated with more skin damage in irradiated mice and low APN levels were associated with more bowel and urinary toxicities and fatigue in human prostate cancer patients. To further investigate the role of APN in radiation damage, we used primary fibroblasts as a model as these cells are sensitive to radiation and are involved in radiation-induced fibrosis. APN treatment significantly improved growth and survival of human fibroblasts and reduced DNA breaks and myofibroblast differentiation caused by irradiation. In addition, APN treatment resulted in increased AMPK phosphorylation and PPAR-α signaling, likely pathways by which APN treatment protects fibroblasts from radiation-induced damage and differentiation.

We then overexpressed APN in vivo and found in pelvically irradiated mice, that APN overexpression significantly reduced skin damage, Th2 skewing, markers of oxidative and nitrosative stress, and collagen deposition in normal tissues exposed to pelvic irradiation. In addition, the adipose tissues from mice receiving APN treatment had significantly less oxidative and morphological damage when exposed to radiation, indicating that APN protects adipose tissue in an autocrine fashion. Finally, the effects of APN overexpression were investigated in a prostate tumor model in combination with radiotherapy. Studies conducted both in vitro and in vivo demonstrate that the addition of APN did not protect the tumor cells from radiation damage. In all experiments, APN treatment resulted in a slight reduction in prostate cancer growth when combined with radiation, although this was not always significant. This data supports the rationale that increasing APN levels during or following radiation treatment would selectively protect normal tissues from radiation-induced damage.

APN has shown anti-tumor properties in a variety of cancers, including prostate cancer [[84], [85], [86], [87], [88], [89]]. A couple of factors may explain why cancer cells are not protected by APN. AMPK, an inhibitor of the mTOR pathway signaling, is activated in normal and tumor cells. However, tumor cells often are heavily reliant on the mTOR pathway for increased growth and survival, while normal cells are not. Thus, when APN activates AMPK, cancer cell growth will be selectively reduced [[90], [91], [92], [93]]. We demonstrated that mTOR activity is reduced in PC-3 cells after treatment with APN. Another reason for the selective outcomes of APN treatment between normal and prostate cancer cells is that prostate cancer cells have less APN receptors than normal cells [80]. Finally, cancer cells are generally more oxidatively stressed than normal cells and have antioxidant pathways, such as, nuclear factor erythroid 2–related factor 2 (NRF2) highly activated [94]. The addition of APN can enhance antioxidant pathways, but if they are already maximally elevated in cancer cells, APN will have little effect on the tumor redox environment. In support of this, our staining of oxidative and nitrosative markers in tumor tissues demonstrates that exogenous APN had no effect on the tumor redox state but did reduce oxidative stress in normal adjacent tissues.

APN is known to have antioxidant, anti-fibrotic, and anti-inflammatory properties yet the details of its mechanism of action are unknown during radiation exposure. In normal cells, AdipoR1 has been shown to phosphorylate AMPK even with knockdown of AdipoR2 [54], while AdipoR2 is necessary to achieve full PPAR-α activation but does not increase AMPK phosphorylation [55]. Despite this division, both pathways may play a role in protection from oxidative and fibrotic radiation damage. Linking antioxidant protection and APN signaling, APN signaling phosphorylates AMPK which activates Unc-51 like autophagy activating kinase (ULK1) to promote autophagy [93]. ULK1 activates vacuole protein sorting 34 (VPS34), which phosphorylates p62 and sequesters Kelch-like ECH-associated protein 1 (Keap1) for autophagic degradation [95]. This allows Nrf2 translocation to the nucleus and transcription of antioxidant defense genes [22,96]. AMPK activation has also been linked to reductions in lung and liver fibrosis [44]. APN signaling through AdipoR2 increased PPAR-α levels, a transcription factor upregulating some antioxidant enzymes (superoxide dismutase 1 and catalase), but notably, inhibits transforming growth factor-beta (TGF-β) and tumor necrosis factor-alpha (TNFα) signaling that promote fibrosis and inflammation respectively [53,[97], [98], [99]]. These data provide a possible link between APN signaling and some of its potential effects, but further study is planned to better understand how isoform, receptor, and cell type protect from radiation toxicities.

Interestingly, Th2 skewing in the spleens of irradiated mice was also reversed with APN treatment. This suggests that APN's protective effects may not only be local but also systemic. The immune system skewing seen in irradiated mice is consistent with other animal radiation studies [70,[100], [101], [102]]. The potentially systemic, long-term effects of radiation treatment need further investigation, which may provide insights into some post-cancer complications, like tumor recurrence due to a more tumor-tolerant immune environment. The enhanced Th2 skewing could also promote a pro-fibrotic environment and the addition of APN may reduce fibrosis through this mechanism as well.

This study has demonstrated that adipose tissue is highly susceptible to radiation damage. Irradiated adipose tissue had smaller adipocytes and increased inflammation, which was consistent with their atrophied appearance at harvest. Adipose tissue also showed signs of persistent oxidative damage to DNA and lipids. APN is made and secreted mainly by adipocytes and we observed that serum APN levels are significantly reduced following radiation to the pelvis in mice and in humans. There are other adipose depots in the body and the bone marrow also contains adipose tissues capable of producing APN. It is surprising that pelvic adipose tissue damage through radiotherapy can create such systemic changes over such long periods of time. It needs to be noted that all patients received androgen deprivation therapy (ADT) before starting radiotherapy. ADT enhances APN levels [65] and we did not have blood collected before ADT treatment to determine how much ADT affected APN levels in our patient cohort. Thus, at baseline, all patients were on ADT so comparisons made at baseline are not driven by differences in ADT. However, the length of ADT varied; therefore, we limited our studies to only patients that remained on ADT for the later time points. To our knowledge this is the first study to demonstrate that pelvic irradiation leads to reduced systemic APN levels in humans.

This striking observation was also identified in a small clinical study conducted by Huang, X. et al., where childhood cancer survivors who received total body irradiation or abdominal irradiation were assessed for metabolic dysfunction approximately 20 years later [70]. The irradiated survivor group showed a reduction in circulating APN levels and increased risk for type 2 diabetes, as compared to patients not receiving radiation. Furthermore, in their gene ontology analysis, IL-4, a cytokine associated with the Th2 response, was the second most enriched gene in the irradiated patient samples versus the control. This is congruent with our clinical findings that radiation reduces APN levels in prostate cancer patients. This also indicates the immune and adipose dysregulation observed in irradiated mice is similar to changes observed in human patients exposed to therapeutic radiation. Thus, our mouse pelvic irradiation model recapitulates dysregulation observed in human radiotherapy patients.

APN levels significantly inversely correlated with bowel, urinary, overall symptom scores and fatigue following radiotherapy. In this cohort, prostate cancer patients that have low levels of plasma APN (<7500 ng/mL) appear to be at risk for developing more radiation toxicities. Therefore, APN levels could serve as a biomarker for the likelihood of developing radiation complications and identify individuals that would benefit most from the use of a radioprotector. A larger study will need to be conducted to conclusively show this as well as determine what the cutoff values of APN would be for patients at risk of developing toxicities from radiotherapy in prostate cancer patients. We speculate that these findings would be applicable to other cancers treated with pelvic irradiation, i.e., anal, colorectal, and endometrial cancers.

There are some limitations to this trial. First, this is a small study with only 65 patients enrolled, a larger cohort should be conducted to determine if these changes are observed in a larger population of patients. Secondly, this study was started right before the COVID pandemic and several patients had missed blood collections and symptom evaluations due to quarantine. Thirdly, the type and length of ADT treatment varied among patients and a larger cohort would allow us to analyze these effects in future studies. Lastly, these findings do not show that APN is protective of radiation toxicities, it demonstrates an inverse correlation of APN to radiation toxicities in prostate cancer patients. The patients with lower APN levels in this study may have other confounding factors such as obesity and other metabolic disorders that could be making these patients more prone to radiation toxicities and APN is just a marker of these other factors.

APN signal transduction is currently being studied as a possible therapeutic target or marker for many pathologies, including cancer, metabolic disorders, and cardiovascular disease [46,85,89,103,104]. However, using APN as a treatment presents several challenges. APN is already in high concentration in the blood making supplementation difficult. APN is a very large protein, especially in its oligomeric forms, making it hard to commercially produce. The predominant APN isoform and/or a specific receptor also may play a major role in its protective effects and needs to be further explored to identify an appropriate target. Stimulating endogenous APN production or activating downstream targets of APN signaling will need more mechanistic study but may have more promise than direct treatment. In addition, we speculate that protecting adipose tissue from oxidative damage during irradiation may maintain normal adipokine secretion and adipocyte metabolism. We have previously reported that MnTE-2-PyP, a superoxide dismutase mimic, protected adipose tissues from irradiation damage [23]. We speculate that the use of vitamin E to inhibit lipid peroxidation in adipose tissues during irradiation may also provide radiation protection and maintain normal metabolic and secretory functions of adipose tissues to maintain APN levels.

For future studies, we want to investigate the expression of AdipoR1 and AdipoR2 receptors after radiation in a chronic setting and determine the levels in normal tissues as compared to tumor. We also want to investigate the effect of radiation on animals that have low APN levels to further demonstrate the role of APN in radiation protection.

This study shows that APN plasma levels are diminished after radiotherapy, APN levels negatively associate with bowel and bladder dysfunction and APN secretion negatively correlates with fatigue in prostate cancer patients. In mice, we show that over expressing APN during radiation therapy protects normal tissues from radiation damage without protecting cancer cells. In addition, we demonstrate the importance of maintaining adipose tissue integrity, the primary source of APN, as it is particularly susceptible to damage resulting in adverse local and systemic effects. The goal of this research is to highlight new avenues of research into the role that adipose tissue plays in radiotherapy.

Funding

NIH R01 CA255618 (ROD), NIHS10OD023447 (ROD), P20 GM104320 (ROD), NIH 1 R01 CA178888 (ROD), NIHR00R015822 (KAD), FY21 NASA Space Grant Fellowship (JAM), Otis Glebe Foundation (MJB).

CRediT authorship contribution statement

Joshua A. McDowell: Writing – original draft, Methodology, Formal analysis. Elizabeth A. Kosmacek: Writing – review & editing, Methodology, Formal analysis, Data curation. Michael J. Baine: Writing – review & editing, Funding acquisition, Conceptualization. Oluwaseun Adebisi: Writing – review & editing, Methodology, Data curation. Cheng Zheng: Writing – review & editing, Formal analysis. Madison M. Bierman: Writing – review & editing, Formal analysis, Data curation. Molly S. Myers: Writing – review & editing. Arpita Chatterjee: Writing – review & editing, Methodology, Data curation. Kia T. Liermann-Wooldrik: Writing – review & editing, Investigation, Data curation. Andrew Lim: Writing – review & editing, Data curation. Kristin A. Dickinson: Writing – review & editing, Funding acquisition, Formal analysis, Data curation, Conceptualization. Rebecca E. Oberley-Deegan: Writing – review & editing, Funding acquisition, Conceptualization.

Declaration of competing interest

None.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.