Forum Article Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/IC
Synchrotron Radiation X‑ray Fluorescence Elemental Mapping in Healthy versus Malignant Prostate Tissues Provides New Insights into the Glucose-Stimulated Zinc Trafficking in the Prostate As Discovered by MRI Veronica Clavijo Jordan,*,†,‡ Alia Al-Ebraheem,§ Kalotina Geraki,∥ Erica Dao,⊥ Andre F. Martins,†,#,∇ Sara Chirayil,† Michael Farquharson,§ and A. Dean Sherry†,#,○ Downloaded via UNIV PARIS-SUD on August 20, 2019 at 11:39:08 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
†
Advanced Imaging Research Center and Department of Radiology, University of Texas Southwestern Medical Center, Dallas, Texas 75390, United States ‡ Athinoula A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Harvard Medical School, Charlestown, Massachusetts 02129, United States § School of Interdisciplinary Science, McMaster University, Hamilton, Ontario L8S 4K1, Canada ∥ Diamond Light Source, Harwell, Didcot OX11 0DE, United Kingdom ⊥ Department of Physics and Astronomy, McMaster University, Hamilton, Ontario L8S 4K1, Canada # Department of Chemistry, University of Texas at Dallas, Richardson, Texas 75080, United States ∇ Werner Siemens Imaging Center, Eberhard Karls University Tuebingen, Tuebingen, 72076, Germany ○ Vitalquan, LLC, Dallas, Texas 75235, United States S Supporting Information *
ABSTRACT: Prostatic zinc content is a known biomarker for discriminating normal healthy tissue from benign prostatic hyperplasia (BPH) and prostate cancer (PCa). Given that zinc content is not readily measured without a tissue biopsy, we have been exploring noninvasive imaging methods to detect these diagnostic differences using a zinc-responsive MRI contrast agent. During imaging studies in mice, we observed that a bolus of glucose stimulates secretion of zinc from the prostate of fasted mice. This discovery allowed the use of a Gd-based zinc sensor to detect differential zinc secretion in regions of healthy versus malignant prostate tissue in a transgenic adenocarcinoma mouse model of PCa. Here, we used a zinc-responsive MRI agent to detect zinc release across the prostate during development of malignancy and confirm the loss of total tissue zinc by synchrotron radiation X-ray fluorescence (μSR-XRF). Quantitative μSR-XRF results show that the lateral lobe of the mouse prostate uniquely accumulates high concentrations of zinc, 1.06 ± 0.08 mM, and that the known loss of zinc content in the prostate is only observed in the lateral lobe during development of PCa. Additionally, we confirm that lesions identified by a loss of zinc secretion indeed represent malignant neoplasia and that the relative zinc concentration in the lesion is reduced to 0.370 ± 0.001 mM. The μSR-XRF data also provided insights into the mechanism of zinc secretion by showing that glucose promotes movement of zinc pools (∼1 mM) from the glandular lumen of the lateral lobe of the mouse prostate into the stromal/smooth muscle surrounding the glands. Co-localization of zinc and gadolinium in the stromal/smooth muscle areas as detected by μSR-XRF confirm that glucose initiates secretion of zinc from intracellular compartments into the extracellular spaces of the gland where it binds to the Gd-based agent and albumin promoting MR image enhancement.
■
INTRODUCTION Metals play a crucial role in many cellular processes and are heterogeneously distributed within the cell and organs.1 In mitochondria, for example, iron−sulfur clusters act as redox catalysts in the electron transport chain and as catalytic sites in TCA cycle enzymes.2 Cellular respiration, free radical detoxification, and cross-linking of collagen and elastin are regulated by copper-containing enzymes, and both iron and © XXXX American Chemical Society
copper play critical roles in homeostasis of reactive oxygen species (ROS).3 In addition to its well-known structural role in zinc finger biochemistry, divalent zinc also serves as a Special Issue: Metals in Biology: From Metallomics to Trafficking Received: April 17, 2019
A
DOI: 10.1021/acs.inorgchem.9b01132 Inorg. Chem. XXXX, XXX, XXX−XXX
Forum Article
Inorganic Chemistry
Figure 1. (A) Schematic to illustrate the mechanism of MRI zinc detection. (B) Current study design and sample preparation. After resection of the organ, 50 and 10 μm sections were cut and mounted for μSR-XRF and hematoxylin and eosin (H&E) staining, respectively.
regulatory messenger ion within cells and between cells4 and is stored and released along with hormones, enzymes, or metabolites released by secretory cells.5 Tissues such as the endocrine and exocrine pancreas, mammary glands, brain, and prostate are known to contain high levels of zinc, and zinc dysregulation is tightly linked to altered pathological states.6−13 Many different types of zinc-targeted MRI and optical sensors have been developed14−19 in an effort to monitor zinc dysregulation, aid in disease diagnosis, or identify therapeutic targets;20 however, few have been proven to work in vivo. One of the most widely studied zinc-containing tissues is the pancreas. In the endocrine pancreas, zinc is packaged with insulin in β-cell granules, which then exocytose into the extracellular space surrounding β-cells in response to an increase in plasma glucose.5 In an effort to image insulin secretion from the intact pancreas in vivo, we designed a Gdbased zinc-responsive agent that binds with serum albumin only in the presence of excess zinc ions and used it to image insulin secretion (via zinc release) from the rodent pancreas.21−23 Given that the pancreas releases insulin and zinc only after stimulation by an increase in blood glucose levels, we refer to this stimulatory process as glucosestimulated zinc secretion (GSZS). This technology has been used to image an increased functional release of insulin and expansion of the mouse pancreas after mice were placed on a high-fat diet for a period of 12 weeks.21 During those studies, we observed by MRI that the prostate also responds to a sudden increase in blood glucose by releasing zinc ions into the surrounding stromal tissue. Given that the prostate contains the highest levels of zinc in the body, and that zinc levels are markedly decreased in prostate cancer (PCa),24 we tested the possibility of using MRI to monitor the zinc status of prostate tissues in anticipation of detecting differential zinc release between normal versus malignant prostate tissues.25 This was
indeed evident in TRAMP mice, a widely studied model of PCa,26 as hypointense regions of the prostate were later shown by histology to correspond to malignant tissue. Although this is consistent with the known decrease in total zinc content in PCa, the question remained: Does the MRI observation simply reflect the total zinc content of prostate tissue or does it reflect the loss of ability of PCa cells to secrete zinc ions in response to a glucose bolus? Therefore, in an effort to further understand these MRI observations, we turned to highresolution synchrotron radiation X-ray fluorescence (μSRXRF) studies of prostate tissue samples from the same animals studied by MRI to obtain elemental maps of zinc and gadolinium before and after exposure to high glucose. X-ray fluorescence (XRF) methods have been used to quantify the distribution of metals in tissues and in various other samples ranging from geology, to electronics, to materials science. This work highlights the advantage of using both MRI and μSRXRF to characterize the distribution and trafficking of zinc in healthy and malignant prostate tissue.
■
RESULTS AND DISCUSSION In this study, we used a combination of MRI to monitor glucose-stimulated zinc secretion (GSZS) in the mouse prostate followed by resection of the prostate, fixation, tissue sectioning, and analysis by μSR-XRF to obtain elemental maps of the tissue distribution of zinc and gadolinium in healthy versus cancerous mouse prostate tissues. For MRI sensing, we used a Gd-DO3A derivative containing a single zinc binding moiety (Figure 1A). This zinc sensor, referred to as GdL2 in a previous publication,22 has a rather modest zinc binding affinity of KD(Zn) = 2.35 μM, as measured by a competitive binding assay in the absence of albumin. Once a single zinc ion binds to GdL2, the resulting complex forms a ternary complex with serum albumin, which results in slowing of molecular B
DOI: 10.1021/acs.inorgchem.9b01132 Inorg. Chem. XXXX, XXX, XXX−XXX
Forum Article
Inorganic Chemistry
Figure 2. In vivo GSZS MRI identification of malignant prostate lesions and μSR-XRF validation of zinc and gadolinium content. (A) T1-weighted gradient echo images of healthy C57BL6 and 21-week-old TRAMP mice prior to and 7 min after receiving a bolus of 0.07 mmol/kg GdL2 plus 2.2 mmol/kg glucose. The inset in the TRAMP mouse image shows a hypo-intense region in the lateral/ventral lobe, consistent with a nascent tumor. (B) Image analysis of ROIs drawn along the prostates of healthy and TRAMP animals. The contrast-to-noise ratio of prostate versus muscle tissue indicates a loss of GSZS in the TRAMP animal (N = 4 healthy, N = 6 TRAMP, *p < 0.05). (C) Representative H&E stains and μSR-XRF images of prostate tissue samples show a poorly differentiated tumor in the lateral lobe and the distribution of zinc, gadolinium, and phosphorus in those same slices. Concentration bars for zinc and gadolinium correspond to absolute quantified values in millimolar; absolute concentrations for P were not obtained due to low energy of the fluorescent signal.
resection and freezing, and preparation of mouse prostate whole mounts on XRF-compatible Ultralene film for μSR-XRF. Injection of glucose and GdL2 resulted in dramatic image enhancement of the prostate in fasted healthy mice (Figure 2A). Image enhancement is not detected without the addition of glucose.25 Although GdL2 has a much lower affinity for zinc than the agent used in our previous publication, which contains two zinc binding arms,25 this result shows that a substantial amount of zinc is secreted from the prostate in response to a sudden increase in glucose, enough to occupy the single binding site on GdL2 and promote detectable complexation with albumin. In 20−23-week-old TRAMP mice, somewhat lower image enhancement is seen from some regions of prostate tissue, plus there are clear hypo-intense regions indicative of reduced zinc flux in those areas. We previously reported that these hypo-intense regions correspond to malignant tissue confirmed by immunohistochemistry.25 A quantitative analysis of these regions, summarized in the boxwhisker plot of Figure 2B, shows that the contrast-to-noise ratio (prostate versus muscle tissues) in the healthy animals at 7 min after GSZS was 15 ± 4.5 versus 8.5 ± 3.7 (p < 0.05) in TRAMP mice. This substantial decrease in CNR is consistent
rotation and an amplification of the longitudinal T1 relaxivity (r1) of water protons at 1.5 T from 4.7 ± 0.1 mM−1 s−1 to 11.1 ± 0.4 mM−1 s−1. Zinc secretion from the prostate is initiated by a bolus injection of glucose in fasted mice followed by intravenous injection of GdL2. Given that imaging agents like GdL2 are thought to remain extracellular after injection, the MR image enhancement detected in the prostate has largely been shown to reflect zinc ions secreted from the prostate initiated by the sudden increase in extracellular glucose.25 Since total tissue zinc content may not necessarily parallel release of zinc as initiated by glucose (GSZS), the present study was designed to measure GSZS by MRI in healthy and TRAMP mouse prostates followed by resection of the prostate and analysis of metal content and distribution by μSR-XRF. Healthy C57BL6 mice and 20−23 week-old TRAMP mice were fasted overnight (at least 12 h) before receiving a bolus of either 2.2 mmol/kg glucose or saline intraperitoneally and 0.07 mmol/kg GdL2 intravenously. As summarized in Figure 1B, T1-weighted MRI scans of the prostate were collected sequentially over ∼7 min followed by removal of the animal from the magnet, prostate C
DOI: 10.1021/acs.inorgchem.9b01132 Inorg. Chem. XXXX, XXX, XXX−XXX
Forum Article
Inorganic Chemistry
Figure 3. Glucose-stimulated movement of zinc pools in the prostate. (A) μSR-XRF of WT C57Bl6 animals without any glucose or GdL2 illustrates the endogenous distribution of zinc. (B) μSR-XRF of WT animals only receiving 0.07 mmol/kg GdL2 shows the distribution of zinc and gadolinium in the prostate. High resolution μSR-XRF shows glands found in the lateral lobe full of zinc-rich secretions. Phosphorus membrane seen in blue delineates the separation between intraglandular zinc and stromal gadolinium. (C) μSR-XRF of WT animals receiving both glucose and GdL2 shows that the distribution of zinc continues to concentrate in the lateral lobe, but at the glandular level, zinc is now secreted from the glands into the stromal space where it comes in contact with gadolinium. (D) μSR-XRF of 20−23-week-old TRAMP receiving glucose and GdL2 illustrates the relative reduction of zinc in the entire gland; at the glandular level, zinc is found predominantly in the smooth muscle surrounding the gland, and not distributed in the stroma as seen in C.
in tumor tissue was 0.370 ± 0.003 mM, while the average zinc content across all other regions was quite variable at 0.40 ± 0.02 mM. Quantitative XRF maps were also measured on prostate tissues from wild-type (WT) mice that had not received glucose or GdL2. The μSR-XRF map in Figure 3A shows that the lateral lobe has much more zinc than the other prostate lobes. A magnified image of the lateral lobe shows that zinc accumulates in the glandular wall (which includes secretory epithelial cells) and the supporting stromal tissue. In animals that received GdL2 but no additional glucose, the XRF images (Figure 3B) show that GdL2 distributes in the surrounding stroma and interstitial spaces but not in the inner luminal, zincrich spaces. The phosphorus maps clearly define the glandular wall that separates zinc from GdL2. Figure 3C shows images from another mouse that received both glucose and GdL2. Here, the glandular lumen is relatively free of zinc, while the stromal and interstitial spaces contain both zinc and gadolinium. These μSR-XRF data confirm the MRI results by directly showing that a bolus of glucose administered to fasted mice stimulates movement of zinc from the glandular
with reduced zinc secretion from the hypointense regions of the TRAMP prostate. Given the known relationship between zinc content in healthy (high zinc) versus cancerous prostate tissue (low zinc),24,27,28 this imaging result could reflect either a loss of ability to secrete zinc ions in cancerous tissues or simply that less zinc is available in the tissue to be secreted. In an effort to differentiate between these possibilities, these tissue samples were frozen by exposure to liquid nitrogenchilled isopentane and prepared for quantitative elemental analyses by μSR-XRF. μSR-XRF images of the prepared prostate tissue slices were collected at the Diamond Light Source, Harwell, UK, by rastering the 50-μm-thick sample with 11 and 8.2 keV beams through the prostate section to obtain elemental maps of zinc, gadolinium, and phosphorus (Figure 2C). The H&E stain shows a poorly differentiated tumor in the lateral lobe of the prostate. The phosphorus map (blue) outlines the phospholipid membranes. The gadolinium map (red) shows that the contrast agent has limited access to the center of the dense tumor core, while the zinc map (green) shows that the zinc content in the tumor core is lower compared to the surrounding tissue. The average zinc content D
DOI: 10.1021/acs.inorgchem.9b01132 Inorg. Chem. XXXX, XXX, XXX−XXX
Forum Article
Inorganic Chemistry
Figure 4. Quantification of element concentration in different prostate lobes. (A) Mouse prostate section illustrating the ventral (yellow), lateral (gray), dorsal (blue), and anterior (green) lobes. Close inspection of histological structures and the decision chart from ref 29 allowed consistent identification of lobular structures. (B) Quantified concentrations of zinc, gadolinium, copper, and iron, categorized by prostate lobe for healthy ̈ WT animals (N = 3), WT mice receiving GdL2 and no glucose (N = 3), WT mice receiving GdL2 and glucose (N = 3), and 20−23-week-old naive TRAMP receiving GdL2 and glucose (N = 10).
we were able to quantify the local metal concentrations in the prostate of animals with or without prior injection of glucose (Figure 5A). The quantitative measures of zinc and gadolinium (Figure 5B) show that glucose effectively stimulates movement of ∼1 mM zinc from the glandular lumen to the stromal compartment (1.5 ± 0.8 mM in the glandular lumen prior to glucose stimulation and 0.5 ± 0.1 mM in the lumen after glucose stimulation, p < 0.05). Furthermore, a significant loss
lumen to the stromal and interstitial spaces where it comes into contact with GdL2. Similarly, μSR-XRF prostate images from TRAMP mice given both glucose and GdL2 showed a similar zinc and gadolinium colocalization in the stromal smooth muscle surrounding the glands, but little zinc in the interstitial spaces (Figure 3D). To quantify the elemental distributions in the respective prostatic lobes, ROIs guided by the adjacent H&E sections were drawn to identify the ventral, lateral, and dorsal lobes (Figure 4A) following previously published guidelines.29 Here, we analyzed zinc, gadolinium, iron, and copper in an effort to determine if other elements play significant roles in the development of PCa in conjunction with zinc. Analyses of these data (Figure 4B) indicate that the lateral lobe of the WT mouse prostate contains the most zinc, 1.06 ± 0.08 mM, compared to the ventral and dorsal lobes, 0.3 ± 0.1 mM and 0.509 ± 0.003, respectively (p = 0.002). For the animals that received GdL2, there was no significant difference in the distribution of the agent across lobes as seen in the quantified gadolinium bar graphs. Furthermore, the known loss of zinc in PCa was found to be localized only in the lateral lobe of the TRAMP mice, decreasing to 0.38 ± 0.02 mM (p-value
