Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Nov 15.
Published in final edited form as: Cancer Res. 2011 Sep 26;71(22):6948–6956. doi: 10.1158/0008-5472.CAN-11-1095

Metabolic signatures imaged in cancer-induced cachexia

Marie-France Penet 1,*, Mayur M Gadiya 1,+, Balaji Krishnamachary 1,+, Sridhar Nimmagadda 1,+, Martin G Pomper 1,2, Dmitri Artemov 1,2, Zaver M Bhujwalla 1,2,*
PMCID: PMC3217079  NIHMSID: NIHMS328410  PMID: 21948967

Abstract

Cancer-induced cachexia is a complex and poorly understood life-threatening syndrome that is characterized by progressive weight loss due to metabolic alterations, depletion of lipid stores and severe loss of skeletal muscle protein. Gaining the ability to non-invasively image the presence or onset of cachexia is important to better treat this condition, to improve the design and optimization of therapeutic strategies, and to detect the responses to such treatments. In this study, we used noninvasive magnetic resonance spectroscopic imaging (MRSI) and [18F] fluorodeoxyglucose (18FDG) positron emission tomography (PET) to identify metabolic signatures typical of cachectic tumors, using this information to determine the types and extents of metabolic changes induced by the onset of cachexia in normal tissues. Cachexia was confirmed by weight loss as well as analyses of muscle tissue and serum. In vivo, cachexia-inducing MAC16 tumors were characterized by higher total choline (tCho) and higher 18FDG uptake compared to histologically similar non-cachectic MAC13 tumors. A profound depletion of the lipid signal was observed in normal tissue of MAC16 tumor bearing mice but not within the tumor tissue itself. High-resolution 1H MR spectroscopy (MRS) confirmed the high tCho level observed in cachectic tumors that occurred due to an increase of free choline and phosphocholine (PC). Higher succinate and lower creatine levels were also detected in cachectic tumors. Taken together, these findings enhance our understanding of cancer’s effect on host organs and tissues as well as promote the development of noninvasive biomarkers for the presence of cachexia and identification of new therapeutic targets.

Keywords: cancer-induced cachexia, magnetic resonance spectroscopic imaging, 18FDG PET, metabolism

INTRODUCTION

Cachexia is a complex metabolic and nutritional syndrome characterized by a massive loss of adipose tissue and skeletal muscle (1). The term cachexia is coined from the Greek words ‘kakos’ and ‘hexis’ for ‘bad condition’ (2). Cachexia is encountered not only in cancer, but also in other life-threatening diseases such as acquired immunodeficiency syndrome, rheumatoid arthritis, chronic obstructive pulmonary disease, and organ failure (1). In cancer patients, the cachectic syndrome is a major cause of morbidity and mortality. Progressive cachexia indicates poor prognosis with a shorter survival time, and ultimately accounts for nearly 20% of all cancer deaths (3, 4). The rate and amount of weight loss are directly related to survival in cancer patients (5). Cachexia occurs with a frequency of 83 to 87% in patients with pancreatic and gastric cancer, 48 to 61% in patients with cancers of the colon, prostate, lung, head and neck, and unfavorable non-Hodgkins lymphoma, and 31 to 40% in patients with breast cancer, sarcomas, leukemia, and favorable subtypes of non-Hodgkins lymphoma (4).

Cachexia involves the secretion of tumor-derived catabolic factors along with host-derived pro-inflammatory cytokines, such as tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), IL-1β and interferon-γ(IFN-γ). Tumor-derived catabolic factors such as lipid-mobilizing factor (LMF) lead to changes in host metabolism and energy expenditure, resulting in cancer cachexia (6). LMF is a toxohormone that up-regulates degradative pathways in adipose tissue, and increases fat metabolism in cachectic animals and humans (6). Glucocorticoids and tumor-derived proteolysis-inducing factor (PIF) also play critical roles in this condition (7). The consequences of the significant weight loss and the massive loss of proteins are multiple and severe, ranging from impairment of immune function to poor outcome of chemotherapy, fatigue, respiratory muscle weakness and markedly reduced quality of life (4, 5).

There is limited understanding of the pathogenesis of this lethal syndrome, and no known cure, since mechanisms underlying its manifestation are not defined clearly enough to design effective therapeutic strategies. As a result therapy is ineffective, and there is an urgent need to identify new targets to develop effective therapies. The ability to detect the onset or presence of cachexia non-invasively, and detect response to treatment, is central to developing new therapies. Significant abnormalities in carbohydrate, lipid and protein metabolism are observed with cachexia, and are a major cause for the associated profound weight loss (5).

Non-invasive clinically translatable techniques such as magnetic resonance spectroscopy (MRS) and positron emission tomography (PET) with their respective metabolic imaging capabilities provide the opportunity to investigate the interaction between cancer cells and the tumor, the ‘tumor macroenvironment’ to identify stages in the cachexia-cascade. With its clinical translatability, relatively high sensitivity and ability to detect metabolites such as lipids, lactate, and total choline (tCho), which consists of free choline (Cho), phosphocholine (PC) and glycerophosphocholine (GPC), 1H MRS imaging (MRSI) is ideally suited to explore the interactions between the tumor and its host in the induction and progression of cachexia. Similarly, [18F] fluorodeoxyglucose (18FDG) positron emission tomography (PET) is frequently used in oncological diagnosis. Here we have identified tumor and normal tissue metabolic signatures associated with the well-established cachexia-inducing murine adenocarcinoma (MAC) 16 tumor model, compared to the histologically similar but non-cachectic MAC13 tumor model.

The presence of cachexia in our experimental model was confirmed by weight loss and an increase in the expression levels of two markers of muscle protein degradation, Atrogin-1 (mouse atrophy gene 1) and MuRF1 (muscle RING finger 1). Atrogin-1 and MuRF1 are two skeletal muscle specific ubiquitin ligases required in protein breakdown (8), and the expression levels of these ligases increases with cachexia, and (5, 8–10). The presence of cachexia was additionally confirmed by the analysis of lipid content in serum from cachectic and non-cachectic tumor bearing mice (11, 12).

In vivo, cachectic tumors were characterized by increased tCho and 18FDG uptake compared to non-cachectic ones, and induced a profound depletion of triglycerides in normal tissue that was not observed with non-cachectic tumors. Cachexia also resulted in increased 18FDG uptake in the brain and lung. An increase of PC and succinate along with a decrease of creatine were detected in high-resolution spectra of cachectic tumor extracts. The higher choline kinase mRNA levels observed in MAC16 cells and tumors were consistent with higher PC observed in tumor extract spectra and the increased tCho observed in vivo. These metabolic patterns may represent new noninvasive biomarkers and targets in the detection, management, and treatment of this condition.

MATERIALS AND METHODS

Cell lines and tumor models

Cachectic MAC16 and non-cachectic MAC13 cell lines were obtained from Dr. Sidransky at Johns Hopkins University with Dr. Tisdale’s permission. Cachectic MAC16 tumors induce significant weight loss in tumor-bearing animals, unlike MAC13 tumors, which although histologically similar, do not alter body weight (13). Cells were cultured in RPMI 1640 with 10% fetal bovine serum. Tumors were generated by inoculating a cell suspension of 2 × 106 cells in 0.05 ml of Hanks balanced salt solution in the flank of severe combined immunodeficient (SCID) male mice. Tumor volumes were calculated from caliper measurements using the equation, volume = (π/6) × a × b × c, where a, b, and c represent three orthogonal axes of the tumor. Since these are murine cancer cell lines, authentication with commercially available kits that are used for human cancer cell lines is not possible. Instead, tumor-induced weight loss, expression of muscle degrading markers, and serum lipid profiles were used to authenticate the cells.

FDG PET imaging

For 18F-fluorodeoxyglucose (FDG) PET imaging, MAC13 and MAC16 tumor-bearing mice were fasted overnight. Mice were injected intravenously (i.v.) with 200 μCi of FDG in a final volume of 200 μl. At 60 minutes post-injection of the radiotracer, a 30-minute static whole-body image was acquired over the mouse body. Images were decay corrected and reconstructed using 2D OSEM (Ordered Subset Expectation Maximization). Data analysis was performed based on regions of interest (ROI) drawn over the tumor and tissues. After accounting for injected dose and body weight, all radioactivity concentration values were converted into Standardized Uptake Values (SUV). Values shown are means of the pixel values (SUVmean). In the case of tumor, to avoid bias from the inclusion of any tumor necrotic areas in the ROI, SUVmean as well as SUVmax (single maximum pixel value within the ROI) were reported. Image analysis was performed using AMIDE (A Medical Image Data Examiner) software (SourceForge).

In vivo 1H MR spectroscopic imaging (MRSI)

Anesthetized mice bearing tumor xenografts were imaged on a 4.7T Bruker Biospec spectrometer (Bruker Biospin Corp., Billerica, MA). Body temperature of the animals was maintained in the magnet by a thermostat-regulated heating pad. To acquire localized proton spectra of the tumor, a home-built solenoid coil was placed around the tumor. Spectra from a 4 mm thick slice were acquired with a field of view (FOV) of 16 mm; a matrix size of 16 × 16 × 1024; 4 scans per phase encode step; number of scans (NS) of 4, an echo time (TE) of 120 ms and a repetition time (TR) of 1 s, using a 2D-CSI sequence with VAPOR water suppression (14). Signals at ~ 3.2 ppm from tCho and at 1.3 ppm from lipids with some contribution from lactate that overlaps with the lipid signal were detected in localized proton spectra. A home-built volume coil was used to acquire whole-body 1H MR spectroscopic images and obtain in vivo cross-sectional images from the tCho and lactate+lipids signal of the mouse. Spectra from a 4 mm thick slice were acquired with a FOV of 32 mm; a matrix size of 32 × 32 × 1024; 4 scans per phase encode step; NS = 4, TE = 120 ms and TR = 1 s. Reference 2D-CSI images of the unsuppressed water signal were acquired with TE = 20 ms and NS = 1, with all other parameters remaining the same.

Quantitative maps of tCho and lactate+lipids were generated from the spectroscopic images using unsuppressed water signal as an internal reference (15). Quantitative metabolic maps were generated using in-house IDL programs.

High-resolution 1H MR spectroscopy of dual phase tissue extracts

High-resolution 1H MR spectroscopy of lipid- and water-soluble fractions of tissue extracts was performed to resolve the components of the tCho signal and the lactate+lipids signal detected in vivo and further characterize the differences observed in vivo. Lipid- and water-soluble fractions were obtained from tumors and muscle tissue using a dual-phase extraction method with methanol/chloroform/water (1/1/1) (16). Briefly, tissues were freeze-clamped and ground to powder. Ice-cold chloroform, followed by ice-cold water, was added and the sample was kept at 4°C overnight for phase separation. Samples were then centrifuged for 30 minutes at 15,000 g at 4°C to separate the phases. The water/methanol phase containing the water-soluble metabolites was treated with chelex (Sigma Chemical Co., St Louis, MO) for 10 minutes on ice to remove divalent cations. Methanol was then removed by rotary evaporation, and the remaining water phase was lyophilized and stored at −20°C. The chloroform phase containing the lipids was dried in a stream of N2 and stored at −20°C. Water-soluble samples were dissolved in 0.5 ml of D2O (Sigma Chemical Co., St Louis, MO) containing 3-(trimethylsilyl) propionic-2, 2, 3, 3,-d4 acid (Sigma Chemical Co., St Louis, MO) as an internal concentration standard (sample pH of 7.4). Lipid samples were dissolved in 0.6 ml of CDCl3/CD3OD (2/1) containing tetramethylsilane as an internal concentration standard (CDCl3 and CD3OD premixed with tetramethylsilane by the manufacturer, Cambridge Isotope Laboratories, Inc.). Fully relaxed 1H MR spectra of the extracts were acquired on a Bruker Avance 500 spectrometer operating at 11.7 T (Bruker BioSpin Corp., Billerica, MA) using a 5-mm HX inverse probe and the following acquisition parameters: 30° flip angle, 6000 Hz sweep width, 12.7 s repetition time, time-domain data points of 32K, and 128 transients (16). Spectra were analyzed using Bruker XWIN-NMR 3.5 software (Bruker BioSpin). Integrals of the metabolites of interest were determined and normalized to the tumor weight. To determine concentrations, peak integration from 1H spectra for all metabolites studied was compared to the internal standard.

Quantitative real-time PCR of cells and tumor extracts

Total RNA was extracted from MAC13 and MAC16 cells and tumors, using Qiagen easy mini kit (Qiagen, CA, USA) as previously described (17). Complementary DNA (cDNA) was synthesized from 1μg of total mRNA using qScript (Quanta Bioscience, MD, USA). Quantitative real-time PCR (q-RT-PCR) was performed using iQ SYBR Green Supermix, gene-specific custom-made primers, and diluted cDNA (1:10) samples as template, in the iCycler RT-PCR detection system (Bio-Rad, CA, USA). The following primers were designed using Beacon Designer software 5.1 (Premier Biosoft International, CA, USA): mChk (NM_013490) forward 5′-CCAGTTCCACATCAGTGTCATCAG-3′, reverse 5′-CTCTCCAGAACCATCGCCTCAG-3′; hypoxanthine phosphoribosyltransferase-1 (HPRT1) (NM_000194) forward 5′-CCTGGCGTCGTGATTAGTGATG-3′, reverse 5′-CAGAGGGCTACAATGTGATGGC-3′ (annealing temperature, 60°C). The expression of the target RNA relative to the housekeeping gene HPRT1 was calculated based on the threshold cycle (Ct) as R = 2−Δ (ΔCt), where ΔCt = Cttarget − Ct HPRT1 and Δ (ΔCt) = ΔCtMAC16 − ΔCtMAC13.

Quantitative real-time PCR of muscle tissue

Total RNA was extracted from the gastrocnemius muscle of mice bearing MAC13 or MAC16 tumors using Trizol (Invitrogen, Carlsbad, USA). Briefly, the muscle tissue was freeze-clamped, ground to powder, and 2 ml of Trizol added per 250 to 400 mg of ground tissue and homogenized. Further steps to purify RNA were carried out following the manufacturer’s recommendation (Invitrogen). cDNA, q-RT-PCR and determination of the fold changes were performed as described previously. The following mouse specific primers at optimized annealing condition were used: Tripartite motif-containing 63 (Trim63/MuRF1, NM_001039048), forward- 5′-AAGACTGAGCTGAGTAACTG-3′, reverse-5′-TAGAGGGTGTCAAACTTCTG-3′, (annealing temperature, 56.1°C); F-box protein 32 (Atrogin-1, NM_026346), forward- 5′-AGAAAGAAAGACATTCAGAACA-3′, reverse-5′-GCTCCTTCGTACTTCCTT-3′, (annealing temperature, 54°C).

Protein isolation and western blot analysis of muscle tissue

Total protein from freeze-clamped muscle tissue was extracted using RIPA buffer fortified with protease inhibitor cocktail (1/500, SIGMA), dithiothreitol (1/1000, 1M stock), phenylmethylsulfonyl fluoride (1/200, 0.2M stock), sodium orthovanadate (1/500, 0.5M stock) and sodium fluoride (1/500, 0.5M stock). About 100 μg of protein was resolved on 10% SDS-PAGE, transferred onto nitrocellulose membranes, and probed with a goat polyclonal antibody (1/1000 in 5% non-fat dry milk) against Atrogin-1 (Everest biotech, Oxfordshire, UK) or rabbit polyclonal (1/1000 in 5% non-fat dry milk) against MuRF1 (Novus Biologicals, Littleton, Co, USA). Appropriate horseradish peroxidase (HRP) conjugated secondary antibody donkey anti-goat (Novus Biologicals, Littleton, Co, USA) or donkey anti-rabbit (GE Healthcare Bio-Sciences, Piscataway, NJ, USA) were used at 1/2000 dilution. A mouse monoclonal antibody against GAPDH (Sigma) at 1/1000 was used as loading control. Immunoblots were developed using SuperSignal West Pico chemiluminescent substrate kit (Pierce Biotechnology, Inc., Rockford, IL, USA).

Serum analysis

Lipid analysis of serum was performed by the Johns Hopkins Phenotyping and Pathology Core. Lipids measured included cholesterol, triglycerides, LDL (low density lipoprotein) and HDL (high density lipoprotein).

Statistical analysis

Values displayed are mean ± SEM. Statistical significance was evaluated using the Student’s t-test; P < 0.05 was considered significant.

RESULTS

Confirming the presence of cachexia

Cachectic MAC16 and non-cachectic MAC13 tumors displayed comparable growth rates and histology. However, as shown in Figure 1a, MAC16 tumors induced significant weight loss, confirming the induction of cachexia. MAC16 and MAC13 tumor-bearing mice that were imaged weighed 22.29 ± 0.89 g and 25.41 ± 0.55 g (P < 0.05) respectively. Tumor volumes at 535.8 ± 44.3 mm3 for the MAC16 and 493.8 ± 40.6 mm3 for MAC13 tumors were not significantly different. The mRNA (Figure 1b) and protein expression levels (Figure 1c) of MuRF1 and Atrogin-1, which are markers of muscle degradation, were significantly higher in muscle tissue of mice bearing MAC16 tumors compared to MAC13 tumors. Serum from cachectic mice was characterized by significantly lower triglycerides and higher LDL (Figure 1d). Total cholesterol and HDL levels were not significantly different.

Figure 1.

Figure 1

Open in a new tab

(a) Normalized weights of MAC13 ( Inline graphic) and MAC16 (■) tumor-bearing mice over time. (n = 10 per group, *P < 0.05). (b) mRNA expression levels of MuRF1 and Atrogin-1 analyzed by RT-PCR in MAC13 ( Inline graphic) and MAC16 (■) muscle tissue. Values represent Mean ± SEM (n = 4, ** P < 0.01, *** P < 0.005). (c) Representative immunoblot showing MuRF1 and Atrogin-1 proteins in muscle tissue obtained from mice bearing from MAC13 or MAC16 tumors. GAPDH was used as loading control (n = 3). (d) Lipid analysis of serum from mice bearing MAC13 or MAC16 tumors. Values represent Mean ± SEM (MAC13, n = 8, MAC16, n = 7; *P < 0.05, **P < 0.01).

Cachectic tumors show higher glucose uptake than non-cachectic tumor and induce increased uptake of glucose in brain and lungs

Representative images shown in Figure 2a demonstrate the increased uptake of 18FDG in MAC16 compared to MAC13 tumors. Quantification of 18FDG uptake confirmed a significant difference between MAC13 and MAC16 tumors (Figure 2b). 18FDG uptake in the brain (Figure 2c) and lung (Figure 2d), but not muscle (Figure 2e), was significantly higher in mice bearing cachectic tumors.

Figure 2.

Figure 2

Open in a new tab

(a) Representative PET images of 18FDG uptake in MAC13 and MAC16 tumors (SUV: standardized uptake value). Quantification of the uptake in (b) tumors, (c) brain, (d) lungs and (e) muscle. Values represent Mean ± SEM (MAC13, n = 8, MAC16, n = 5; * P < 0.005).

Cachectic tumors have altered total choline levels in vivo

Representative high-resolution T1-weighted images of a MAC13 (upper panel) and MAC16 (lower panel) presented in Figure 3a demonstrate that the MAC16 (lower panel) tumor had significantly higher tCho compared to the MAC13 tumor (upper panel), as evident in the tCho maps shown in Figure 3b. Representative lactate+lipids images from these tumors are shown in Figure 3c and demonstrate that, although the lactate+lipids content was not different between the MAC13 (upper panel) and the MAC16 (lower panel) tumor, the peripheral signal, most likely from subcutaneous lipids, was lower around the MAC16 tumor.

Figure 3.

Figure 3

Open in a new tab

Representative (a) T1-weighted images, (b) tCho maps and (c) lactate+lipids maps of a MAC13 (upper panel) and a MAC16 (lower panel) tumor. T1-weighted images were acquired from the corresponding 4 mm slice used for MRSI using a spin-echo sequence with an echo time of 10 ms, a repetition time of 500 ms, and an in-plane spatial resolution of 62.5 μm. Total choline maps and lactate+lipids maps were generated from the MRSI data and normalized to the water signal to display concentrations in mM units. (d) In vivo tCho concentrations in MAC13 and MAC16 tumors. (e) In vivo lactate+lipids concentration within MAC13 and MAC16 tumors. (f) In vivo concentration of subcutaneous lipids surrounding MAC13 and MAC16 tumors. Values represent Mean ± SEM (n = 10 per group, * P < 0.05).

Quantitative tCho and lactate+lipids data acquired from multiple animals are summarized in Figures 3d-f. A significant increase of tCho was observed in MAC16 compared to MAC13 tumors (Figure 3d). There was no significant difference in lactate+lipids obtained from within the tumor (Figure 3e), but a significant decrease of lactate+lipids around the tumor periphery in MAC16 compared to MAC13 tumors (Figure 3f). The lactate+lipids signal around the periphery of the tumor was from subcutaneous fatty tissue that is found between the tumor and the fold of the skin, which decreased as the lipids decreased in normal tissue of cachectic mice, as discussed subsequently.

Cachectic tumor induces lipid depletion in normal tissue

To determine the effect of cachectic tumors on normal tissue tCho and lactate+lipids, we performed in vivo 1H MRSI of the mouse whole body using a volume coil. Representative high-resolution T1-weighted images of a MAC13 (upper panel) and MAC16 (lower panel) are presented in Figure 4a to identify the cross-sectional area from which the metabolite images were obtained. Representative lactate+lipids images obtained from this cross-sectional slice are shown in Figure 4b and demonstrate the profound depletion of lactate+lipids signal in normal tissue in MAC16 tumor-bearing mice (lower panel) compared to MAC13 tumor-bearing mice (upper panel), as evident in Figure 4c that displays the lactate+lipids signal overlaid with the spin-density images.

Figure 4.

Figure 4

Open in a new tab

(a) Cross-sectional T1-weighted images, (b) cross-sectional lactate+lipids maps, and (c) merged images from (a) and (b) of MAC13 (upper panel) and MAC16 (lower panel) tumor-bearing mice. T1-weighted images were acquired from the corresponding 4 mm slice used for MRSI using a spin-echo sequence with an echo time of 10 ms, a repetition time of 500 ms, and an in-plane spatial resolution of 125 μm. Lipid maps were generated from MRSI data and normalized to the water signal. Volumes were comparable for the MAC13 (540 mm3) and MAC16 (545 mm3) tumors. (d) In vivo lactate+lipids concentration in MAC13 ( Inline graphic) and MAC16 (■) tumor-bearing mice (n = 6 per group). Values were obtained from a 4 mm axial slice using a volume coil. Values represent Mean ± SEM. * P < 0.05.

Quantitative lactate+lipids data acquired from multiple animals are summarized in Figure 4d. A significant decrease of lactate+lipids was observed in the normal tissue of MAC16 tumor-bearing mice compared to MAC13 tumor-bearing mice that was subsequently identified to be due to lipid depletion. No significant differences in tCho were observed in the normal tissues of MAC16 and MAC13 tumor-bearing mice.

Tumor extract studies confirmed in vivo data

To characterize better the observations made in vivo, we performed high-resolution proton spectroscopy of water-soluble and lipid extracts of tumors. The tCho peak observed in vivo consists of free choline, PC and GPC that can be identified in high-resolution spectra of water-soluble extracts. Similarly, contributions from lactate versus lipids to the lactate+lipids signal observed in vivo can be resolved since lactate is separated into the water-soluble phase and the lipids are separated into the lipid phase of the tumor extracts. Representative spectra from water-soluble tumor extracts are shown in Figure 5a and b, and demonstrate significantly higher PC and free choline in MAC16 (Figure 5b) tumors compared to MAC13 tumors (Figure 5a), but no difference in lactate. Quantitative data acquired from multiple tumors are summarized in Figure 5c and demonstrate a significant increase of free choline, PC, and total choline that confirmed the in vivo data.

Figure 5.

Figure 5

Open in a new tab

Representative 1H MR high-resolution spectra of water-soluble (a) MAC13 and (b) MAC16 tumor extracts. Cho (free choline), PC (phosphocholine), GPC (glycerophosphocholine). (c) Concentration of metabolites (Cho, PC, GPC, total choline = Cho+PC+GPC, succinate and creatine) in MAC13 ( Inline graphic) and MAC16 (■) tumors (n = 6 per group, * P < 0.05, ** P < 0.001). Values represent Mean ± SEM.

Additionally, a significant increase of succinate and a significant decrease of creatine were observed in the water-soluble tumor extracts of MAC16 compared to MAC13 tumors (Figure 5c). There were no significant differences in phosphocreatine, alanine or acetate (data not shown) between MAC16 and MAC13 tumors. Lipid signals in the lipid-phase extracts were not significantly different between MAC16 and MAC13 tumors again confirming the in vivo data (data not shown).

To identify further the molecular mechanism underlying the increase of PC we performed quantitative RT-PCR of choline kinase, the enzyme that phosphorylates free choline to PC. We found a significant increase of choline kinase mRNA in MAC16 cells (Figure 6a) and tumors (Figure 6b).

Figure 6.

Figure 6

Open in a new tab

mRNA levels of Chk analyzed by RT-PCR in MAC13 ( Inline graphic) and MAC16 (■) (a) cell and (b) tumor extracts. Values represent Mean ± SEM (n = 5 per group for the cell extract, n = 8 per group for the tumor extracts, * P < 0.05, *** P < 0.005).

Cachexia affects lipid content of muscle

The profound depletion of lipids in normal tissue observed in vivo was confirmed in high-resolution spectra of extracted muscle tissue. Representative lipid spectra from muscle tissue extract obtained from MAC13 and MAC16 tumor bearing mice are shown in Figure 7a and b respectively. In the spectra, the different peaks are assigned to the protons signals of different groups of fatty acids, methyl (-CH3, 0.9 ppm), aliphatic methylene (-CH2, 1.3 ppm), carboxyl (-OOC-CH2-CH2-, 1.47 ppm; -OOC-CH2, 2.4 ppm), allyl methylene (=CH-CH2, 1.92 ppm), vinyl (=CH-, 5.3 ppm) (18–20). Signals from saturated lipids are visible at 1.3 and 0.9 ppm, poly-unsaturated lipids at 1.92 and 5.3 ppm and the carboxyl group attached to fatty acids at 1.47 and 2.4 ppm. Quantitative data summarized from multiple animals are shown in Figure 7c and demonstrate the significant decrease of lipids induced in the muscle by cachexia. Signals from all of these groups, but especially the carboxyl and methyl groups, decreased in the muscle tissue extracted from cachectic mice.

Figure 7.

Figure 7

Open in a new tab

Representative 1H MR high-resolution spectra of lipid-soluble muscle extracts obtained from (a) MAC13 and (b) MAC16 tumor bearing mice. The chemical shift reference peak is tetramethylsilane that is also used as an internal concentration standard. (c) Quantification of lipids in arbitrary units from lipid-soluble extracts of muscle tissue from MAC13 ( Inline graphic) and MAC16 (■) tumor bearing mice (n = 3 per group). Values represent Mean ± SEM. * P < 0.05.

DISCUSSION

Here we have identified, with noninvasive imaging, clinically translatable metabolic signatures in tumors and normal tissue that potentially detect the presence of cachexia. Cachectic MAC16 tumors displayed increased tCho and 18FDG uptake compared to non-cachectic MAC13 tumors. A profound depletion of the lipid signal was observed in normal tissue of mice bearing cachexia-inducing MAC16 tumors, in addition to increased 18FDG uptake in the brain and lungs. High-resolution spectroscopy of tissue extracts identified additional metabolite signatures, such as increased PC and succinate, and reduced creatine, which may also provide targets to reverse or inhibit the cachexia cascade.

The presence of cachexia induced by MAC16 tumors was confirmed by the significant weight loss, the increase of MuRF-1 and Atrogin-1 mRNA and protein expression in muscle tissue, and a significant decrease of triglycerides and increase of LDL in serum of MAC16 tumor bearing mice. MuRF-1 and Atrogin-1 have been identified as muscle protein breakdown markers that increase with cachexia (10). Atrogin-1 and MuRF1 are skeletal muscle specific ubiquitin ligases involved in the ubiquitin proteasome pathway and essential in the development of muscle atrophy (8, 9). Decreased serum triglycerides and increased LDL have also been previously associated with cachexia (11, 12). Collectively these data confirmed the induction of cachexia in MAC16 tumor-bearing mice.

Elevated tCho is characteristic of malignant tumors compared to normal tissue, but here we observed a two-fold increase of tCho in cachectic MAC16 tumors compared to non-cachectic MAC13 tumors, suggesting that tumors with higher than usual tCho may pose a risk for the induction of cachexia. The increased tCho was identified to be due to significantly higher PC and free choline. An increase of cellular PC and choline-containing compounds has been consistently observed in cancer cells and tissue (21–23), and is closely related to malignant transformation, invasion, and metastasis (24, 25). Elevated PC and tCho have been previously detected by MRS studies in human lung, colon, prostate, and breast tumors, as well as in derived epithelial cell lines (21, 23). PC is a precursor as well as a metabolic breakdown product of the major membrane component phosphatidylcholine. Here we have identified a previously unknown association between choline metabolism and cachexia. We found that mRNA for choline kinase, a cytosolic enzyme that catalyzes the phosphorylation of choline to PC by ATP in the presence of magnesium, was significantly elevated in MAC16 cells and tumors, identifying it as a potential novel target to investigate in the treatment of cachexia.

Increased 18FDG uptake is used to detect metabolically active cancers. 18FDG is taken up by the cells via glucose transporters and phosphorylated by the first enzyme of the glycolysis, the hexokinase. 18FDG PET can be used in vivo to identify tumor tissues, since cancer cells exhibit a higher glycolytic activity, leading to a higher uptake of 18FDG (26). Here we found that similar to tCho, 18FDG uptake was more than two-fold higher in cachectic MAC16 tumors compared to non-cachectic MAC13 tumors, suggesting that exceedingly high 18FDG uptake may be another risk factor for the induction or presence of cachexia. Although high 18FDG uptake is usually associated with increased glycolytic activity (26, 27), we did not observe a significant difference in lactate levels between MAC13 and MAC16 tumors. Interestingly we did detect a higher level of succinate in the cachectic tumors. Since succinate is produced in the tricarboxylic acid cycle (TCA) cycle from α-ketoglutarate (28), increased flux through the TCA cycle may explain a higher uptake of glucose without an increase of lactate.

Subcutaneous lipids located around the tumor were significantly reduced in the cachectic tumor-bearing mice and lipids in surrounding normal tissue and the skeletal muscle of MAC16 tumor-bearing mice were depleted. It has been previously shown that increased adipocyte lipolysis, rather than decrease in synthesis, is responsible for the loss of fat mass with cachexia (29). Moreover acetyl-CoA required for the TCA cycle can be obtained from degradation of lipids. The decrease in lipids observed in cachectic mice may arise from an increase in the flux through the TCA cycle, which is also consistent with the increase of succinate. Depletion of lipids in the muscle can be easily detected clinically with 1H MRS, and may be of value in detecting the presence or onset of cachexia or in evaluating the effect of treatments arresting or reversing this condition.

The search for noninvasive biomarkers of cachexia has been ongoing for several years. One 1H MRS-based search for cachexia biomarkers in the sera of C26 adenocarcinoma-bearing mice identified very low-density lipoproteins along with hypoglycemia as serum markers (11). An assessment of hepatic and skeletal muscle bioenergetic status using 31P MRS demonstrated that progressive cancer cachexia was associated with depletion of energy stores in liver and skeletal muscle (30). Increased hepatic Pi/ATP ratio was detected by 31P MRS, and occurred early in the disease process (30). The liver is a central organ in the pathogenesis of cachexia (31). Increased Pi and a rapid removal of phosphomonoesters were identified with 31P MRS in a rat cancer model of cachexia. Cachexia also induced enhanced hepatic gluconeogenic activity (32). While the use of 31P MRS is less practical for clinical applications due to the poor sensitivity of detection limiting spatial localization, detection of global changes in liver metabolism with 31P MRS may provide additional specificity in detecting the onset or presence of cachexia.

In conclusion we have, in a preclinical model, identified previously unknown clinically translatable tumor and normal tissue metabolic signatures characteristic of cachexia. The ease of translating these signatures to the clinic should facilitate validating these metabolites as potential biomarkers of cachexia. In addition the metabolic changes identified here provide new targets to exploit in the treatment of cachexia. Future studies with noninvasive imaging will allow us to establish a sequence of events to identify the most lethal metabolic aspects of the cascade. Once critical single or multiple cachectic switches or cascades are identified, molecular targeted image-guided approaches such as multiple siRNA targeting or pharmaceutical interventions can be designed and developed to down-regulate or knockout effector molecules, pathways or enzymes controlling the cascade.

Acknowledgments

We thank Mr. Gary Cromwell for maintaining the cell lines and inoculating tumors, and Ms. Flonné Wildes, Mr. James J. Fox and Mr. Gilbert Green for technical assistance. We acknowledge useful discussions with Drs. Paul Winnard and Michael Jacobs. This work was supported by R21 CA140904 and P50 CA103175.

ABBREVIATIONS

AMIDE

A Medical Image Data Examiner

Atrogin-1

mouse atrophy gene 1

Cho

free choline

18FDG

[18F] fluorodeoxyglucose

FOV

field of view

GPC

glycerophosphocholine

HDL

high density lipoprotein

HPRT1

hypoxanthine phosphoribosyltransferase-1

IL-

interleukin-

LDL

low density lipoprotein

LMF

lipid-mobilizing factor

MAC

murine adenocarcinoma

MRS

magnetic resonance spectroscopy

MRSI

magnetic resonance spectroscopic imaging

MuRF1

muscle RING finger 1

NS

number of scans

OSEM

Ordered Subset Expectation Maximization

PC

phosphocholine

PET

positron emission tomography

PIF

proteolysis-inducing factor

q-RT-PCR

quantitative real-time PCR

ROI

regions of interest

SCID

severe combined immunodeficient

SUV

standard uptake value

tCho

total choline

TE

echo time

TNF-α

tumor necrosis factor-α

TR

repetition time

LIST OF REFERENCES

  • 1.Tisdale MJ. Cancer cachexia. Langenbecks Arch Surg. 2004;389:299–305. doi: 10.1007/s00423-004-0486-7. [DOI] [PubMed] [Google Scholar]
  • 2.Inui A. Cancer anorexia-cachexia syndrome: current issues in research and management. CA Cancer J Clin. 2002;52:72–91. doi: 10.3322/canjclin.52.2.72. [DOI] [PubMed] [Google Scholar]
  • 3.Tisdale MJ, Dhesi JK. Inhibition of weight loss by omega-3 fatty acids in an experimental cachexia model. Cancer Res. 1990;50:5022–6. [PubMed] [Google Scholar]
  • 4.Loberg RD, Bradley DA, Tomlins SA, Chinnaiyan AM, Pienta KJ. The lethal phenotype of cancer: the molecular basis of death due to malignancy. CA Cancer J Clin. 2007;57:225–41. doi: 10.3322/canjclin.57.4.225. [DOI] [PubMed] [Google Scholar]
  • 5.Tisdale MJ. Mechanisms of cancer cachexia. Physiol Rev. 2009;89:381–410. doi: 10.1152/physrev.00016.2008. [DOI] [PubMed] [Google Scholar]
  • 6.Cannon T, Couch M, Yin X, Guttridge D, Lai V, Shores C. Comparison of animal models for head and neck cancer cachexia. Laryngoscope. 2007;117:2152–8. doi: 10.1097/MLG.0b013e3181453658. [DOI] [PubMed] [Google Scholar]
  • 7.Todorov P, Cariuk P, McDevitt T, Coles B, Fearon K, Tisdale M. Characterization of a cancer cachectic factor. Nature. 1996;379:739–42. doi: 10.1038/379739a0. [DOI] [PubMed] [Google Scholar]
  • 8.Bodine SC, Latres E, Baumhueter S, Lai VK, Nunez L, Clarke BA, et al. Identification of ubiquitin ligases required for skeletal muscle atrophy. Science. 2001;294:1704–8. doi: 10.1126/science.1065874. [DOI] [PubMed] [Google Scholar]
  • 9.Siddiqui RA, Hassan S, Harvey KA, Rasool T, Das T, Mukerji P, et al. Attenuation of proteolysis and muscle wasting by curcumin c3 complex in MAC16 colon tumour-bearing mice. Br J Nutr. 2009;102:967–75. doi: 10.1017/S0007114509345250. [DOI] [PubMed] [Google Scholar]
  • 10.Peters SJ, van Helvoort A, Kegler D, Argiles JM, Luiking YC, Laviano A, et al. Dose-dependent effects of leucine supplementation on preservation of muscle mass in cancer cachectic mice. Oncol Rep. 2011;26:247–54. doi: 10.3892/or.2011.1269. [DOI] [PubMed] [Google Scholar]
  • 11.O’Connell TM, Ardeshirpour F, Asher SA, Winnike JH, Yin X, George J, et al. Metabolomic analysis of cancer cachexia reveals distinct lipid and glucose alterations. Metabolomics. 2008;4:216 – 25. [Google Scholar]
  • 12.Mahony SM, Beck SA, Tisdale MJ. Comparison of weight loss induced by recombinant tumour necrosis factor with that produced by a cachexia-inducing tumour. Br J Cancer. 1988;57:385–9. doi: 10.1038/bjc.1988.87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Eley HL, McDonald PS, Russell ST, Tisdale MJ. Inhibition of activation of dsRNA-dependent protein kinase and tumour growth inhibition. Cancer Chemother Pharmacol. 2009;63:651–9. doi: 10.1007/s00280-008-0782-y. [DOI] [PubMed] [Google Scholar]
  • 14.Tkac I, Starcuk Z, Choi IY, Gruetter R. In vivo 1H NMR spectroscopy of rat brain at 1 ms echo time. Magn Reson Med. 1999;41:649–56. doi: 10.1002/(sici)1522-2594(199904)41:4<649::aid-mrm2>3.0.co;2-g. [DOI] [PubMed] [Google Scholar]
  • 15.Bolan PJ, Meisamy S, Baker EH, Lin J, Emory T, Nelson M, et al. In vivo quantification of choline compounds in the breast with 1H MR spectroscopy. Magn Reson Med. 2003;50:1134–43. doi: 10.1002/mrm.10654. [DOI] [PubMed] [Google Scholar]
  • 16.Glunde K, Raman V, Mori N, Bhujwalla ZM. RNA interference-mediated choline kinase suppression in breast cancer cells induces differentiation and reduces proliferation. Cancer Res. 2005;65:11034–43. doi: 10.1158/0008-5472.CAN-05-1807. [DOI] [PubMed] [Google Scholar]
  • 17.Krishnamachary B, Glunde K, Wildes F, Mori N, Takagi T, Raman V, et al. Noninvasive detection of lentiviral-mediated choline kinase targeting in a human breast cancer xenograft. Cancer Res. 2009;69:3464–71. doi: 10.1158/0008-5472.CAN-08-4120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Pearlman JD, Zajicek J, Merickel MB, Carman CS, Ayers CR, Brookeman JR, et al. High-resolution 1H NMR spectral signature from human atheroma. Magn Reson Med. 1988;7:262–79. doi: 10.1002/mrm.1910070303. [DOI] [PubMed] [Google Scholar]
  • 19.Zietkowski D, Davidson RL, Eykyn TR, De Silva SS, Desouza NM, Payne GS. Detection of cancer in cervical tissue biopsies using mobile lipid resonances measured with diffusion-weighted (1)H magnetic resonance spectroscopy. NMR Biomed. 2010;23:382–90. doi: 10.1002/nbm.1472. [DOI] [PubMed] [Google Scholar]
  • 20.Wallace JC, Raaphorst GP, Somorjai RL, Ng CE, Fung Kee Fung M, Senterman M, et al. Classification of 1H MR spectra of biopsies from untreated and recurrent ovarian cancer using linear discriminant analysis. Magn Reson Med. 1997;38:569–76. doi: 10.1002/mrm.1910380411. [DOI] [PubMed] [Google Scholar]
  • 21.Glunde K, Jacobs MA, Bhujwalla ZM. Choline metabolism in cancer: implications for diagnosis and therapy. Expert Rev Mol Diagn. 2006;6:821–9. doi: 10.1586/14737159.6.6.821. [DOI] [PubMed] [Google Scholar]
  • 22.Ramirez de Molina A, Gutierrez R, Ramos MA, Silva JM, Silva J, Bonilla F, et al. Increased choline kinase activity in human breast carcinomas: clinical evidence for a potential novel antitumor strategy. Oncogene. 2002;21:4317–22. doi: 10.1038/sj.onc.1205556. [DOI] [PubMed] [Google Scholar]
  • 23.Ramirez de Molina A, Rodriguez-Gonzalez A, Gutierrez R, Martinez-Pineiro L, Sanchez J, Bonilla F, et al. Overexpression of choline kinase is a frequent feature in human tumor-derived cell lines and in lung, prostate, and colorectal human cancers. Biochem Biophys Res Commun. 2002;296:580–3. doi: 10.1016/s0006-291x(02)00920-8. [DOI] [PubMed] [Google Scholar]
  • 24.Katz-Brull R, Lavin PT, Lenkinski RE. Clinical utility of proton magnetic resonance spectroscopy in characterizing breast lesions. J Natl Cancer Inst. 2002;94:1197–203. doi: 10.1093/jnci/94.16.1197. [DOI] [PubMed] [Google Scholar]
  • 25.Negendank W. Studies of human tumors by MRS: a review. NMR Biomed. 1992;5:303–24. doi: 10.1002/nbm.1940050518. [DOI] [PubMed] [Google Scholar]
  • 26.Gambhir SS. Molecular imaging of cancer with positron emission tomography. Nat Rev Cancer. 2002;2:683–93. doi: 10.1038/nrc882. [DOI] [PubMed] [Google Scholar]
  • 27.Gatenby RA, Gillies RJ. Glycolysis in cancer: a potential target for therapy. Int J Biochem Cell Biol. 2007;39:1358–66. doi: 10.1016/j.biocel.2007.03.021. [DOI] [PubMed] [Google Scholar]
  • 28.Perrine SA, Michaels MS, Ghoddoussi F, Hyde EM, Tancer ME, Galloway MP. Cardiac effects of MDMA on the metabolic profile determined with 1H-magnetic resonance spectroscopy in the rat. NMR Biomed. 2009;22:419–25. doi: 10.1002/nbm.1352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Ryden M, Agustsson T, Laurencikiene J, Britton T, Sjolin E, Isaksson B, et al. Lipolysis--not inflammation, cell death, or lipogenesis--is involved in adipose tissue loss in cancer cachexia. Cancer. 2008;113:1695–704. doi: 10.1002/cncr.23802. [DOI] [PubMed] [Google Scholar]
  • 30.Schneeberger AL, Thompson RT, Driedger AA, Finley RJ, Inculet RI. Effect of cancer on the in vivo energy state of rat liver and skeletal muscle. Cancer Res. 1989;49:1160–4. [PubMed] [Google Scholar]
  • 31.Couch M, Lai V, Cannon T, Guttridge D, Zanation A, George J, et al. Cancer cachexia syndrome in head and neck cancer patients: part I. Diagnosis, impact on quality of life and survival, and treatment. Head Neck. 2007;29:401–11. doi: 10.1002/hed.20447. [DOI] [PubMed] [Google Scholar]
  • 32.Gehman KE, Inculet RI, Brauer M, Marsh GD, Driedger AA, Thompson RT. Early detection of cancer cachexia in the rat using 31P magnetic resonance spectroscopy of the liver and a fructose stress test. NMR Biomed. 1996;9:271–5. doi: 10.1002/(SICI)1099-1492(199609)9:6<271::AID-NBM421>3.0.CO;2-8. [DOI] [PubMed] [Google Scholar]

RESOURCES