Cancer Research Annual Meeting 2010 Jordan
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Griffin, J. L.
Right arrow Articles by Kauppinen, R. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Griffin, J. L.
Right arrow Articles by Kauppinen, R. A.
[Cancer Research 63, 3195-3201, June 15, 2003]
© 2003 American Association for Cancer Research

Experimental Therapeutics

Assignment of 1H Nuclear Magnetic Resonance Visible Polyunsaturated Fatty Acids in BT4C Gliomas Undergoing Ganciclovir-Thymidine Kinase Gene Therapy-induced Programmed Cell Death1

Julian L. Griffin, Kimmo K. Lehtimäki, Piia K. Valonen, Olli H. J. Gröhn, Mikko I. Kettunen, Seppo Ylä-Herttuala, Asla Pitkänen, Jeremy K. Nicholson and Risto A. Kauppinen2

Department of Biological Chemistry, Faculty of Medicine, Imperial College of Science, Technology and Medicine, London, United Kingdom [J. L. G., J. K. N.]; Department of Biomedical NMR and National Bio-NMR Facility [K. K. L., P. K. V., O. H. J. G., M. I. K.], Department of Biotechnology and Molecular Medicine [S. Y-H.], and Department of Neurobiology [A. P.], A. I. Virtanen Institute for Molecular Sciences, University of Kuopio, Kuopio, Finland; School of Biological Sciences, University of Manchester, Manchester, United Kingdom [R. A. K.]


   ABSTRACT
Top
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Polyunsaturated fatty acids (PUFAs), as detected by 1H nuclear magnetic resonance (NMR) spectroscopy, accumulate into BT4C glioma during ganciclovir-thymidine kinase gene therapy-induced programmed cell death (PCD). In this study, we have quantified the 1H NMR visible lipids in vivo and characterized their biophysical and biochemical nature in these tumors during PCD both ex vivo and in vitro. Concentrations of 1H NMR-detectable PUFAs increased 3-fold with pattern recognition identifying CH = CH and CH = CHCH2CH = CH as the most significant in monitoring the dynamics of PCD. The increase in PUFAs was equivalent to 70% of that in CH2CH2CH2-saturated lipid peak at 1.3 ppm. Ex vivo tumor samples, obtained from in situ funnel frozen tumors, showed very similar macromolecular peaks, as studied using high-resolution magic angle spinning 1H NMR at 14.1 T, to those detected in vivo at 4.7 T. Line widths of lipid peaks were not influenced by the spin rate within the range of 1–9 kHz or temperature between 277 and 293 K, showing high degree of 1H NMR detection of these peaks in vivo. These biophysical results additionally corroborate the idea that cytoplasmic lipid vesicles are the source of 1H NMR lipid signals. Two-dimensional 1H NMR ex vivo and tumor lipid extracts in vitro showed that the PUFA signals are in the same chemical compounds and consist of largely 18:1 and 18:2 lipids. Furthermore, it is suggested that the 1H NMR lipids detected during PCD arise from cell constituent breakdown products forming lipid vesicles into dying cells.


   INTRODUCTION
Top
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Because of the large dipolar couplings experienced by membrane lipids, 1H NMR3 detects a special class of cell lipids in vivo by revealing sharp resonances from lipids that are present in intracellular lipid vesicles (or lipid bodies) from the cytosol (1, 2, 3) . Outside adipose and muscle tissues, the lipid resonances are particularly prominent both in proliferating and malignant cells (4 , 5) , as well as in a variety of tumors (6) . In malignant cells and tumors, appearance of a lipid peak at 1.3 ppm from CH2CH2CH2 moieties has commonly been associated with ongoing cell death processes (7, 8, 9) . Evidence has been provided that in cell cultures in vitro, PCD leads to accumulation of CH2CH2CH2 lipid groups, whereas necrosis is associated with no signs of 1H NMR lipid retention (7) . In contrast, studies of tumor tissues indicate that presence of this lipid peak coincides with necrotic histopathology (10 , 11) . In rat BT4C glioma, strong 1H NMR lipid peaks are detected in the proliferation phase (1) , and interestingly, ganciclovir-thymidine kinase gene therapy-induced PCD causes accumulation of PUFA (12) . Importantly, in the rat glioma, histological and immunohistochemical analyses reveal no signs of necrotic process during tumor eradication. This suggests that alterations in the degree of lipid saturation, in association with their accumulation, occur in the 1H NMR visible lipids during PCD.

Line broadening of resonances detected in vivo is a result of a number of physical processes, including dipole-dipole coupling, chemical shift anisotropy, and macroscopic and microscopic susceptibility effects. However, these effects can be reduced dramatically by spinning samples at the so-called magic angle (cos{theta} = {surd}1/3; Ref. 13 ), producing high resolution spectra comparable with those obtainable using solution state NMR spectroscopy. For intact tissues, this is further aided by the semi-liquid nature of the cytosol, allowing the ready detection of low molecular weight metabolites (14, 15, 16) . HRMAS 1H NMR spectroscopy is highly effective at characterizing different brain pathologies (17) , liposarcomas (15 , 18) , and prostate cancer (16) , largely as a result of the information contained within the lipid profile of the tissues. However, the spinning speeds that are routinely used in HRMAS 1H NMR spectroscopy (3–6 kHz) are too slow to spin out dipolar couplings in cell membranes (19) , so it is assumed that these lipids are nonmembrane in nature.

In this study, we have carried out in vivo and complementary ex vivo high-resolution 1H NMR spectroscopy of lipids in glioma tissue, addressed the level of polyunsaturated groups in the lipids, and quantified the contributions of individual lipid classes to the NMR-detectable unsaturated/saturated acyl groups. Our data show that there is a distinct temporal progression in terms of metabolic profile of the glioma tissues and cells after PCD, and the PUFA resonance arises from a nonmembrane source of lipids with ADCs consistent with the source being cytosolic lipid droplets. Furthermore, these PUFA resonances arise from mixtures of resonances associated largely with 18:1 and 18:2 lipid moieties.


   MATERIALS AND METHODS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
Animals and Tissue Sampling.
BT4C gliomas transfected with viral HSV-tk gene were induced by implanting 104 HSV-tk+ cells in 5 µl of Optimem to a depth of 2.5 mm into the corpus callosum of female BDIX rats (n = 34) weighing 190–250 g as described previously (20) . Rats in the treatment group were injected with ganciclovir (25 mg/kg, i.p., twice daily for 8 days) for the duration of the study. Untreated tumor-bearing animals served as controls. Rats were funnel frozen in situ (21) , and tumors were dissected frozen on dry ice for sampling for ex vivo NMR spectroscopy and standard lipid extraction procedure (22) . All animal experiments were performed according to the guidelines approved by the Ethical Committee of the National Laboratory Animal Center (Kuopio, Finland).

NMR Procedures.
For MRI, rats were anesthetized with 0.8–1.0% halothane in 7/3 N2O/O2, and the core temperature of animals was monitored and maintained close to 37°C using a heated water blanket. MRI was performed in a horizontal 4.7-T magnet (Magnex Scientific Ltd., Abingdon, United Kingdom), equipped with actively shielded field gradients (Magnex Scientific) interfaced to Varian UNITYINOVA (Varian, Inc., Palo Alto, CA). A quadrature surface coil (Highfield Imaging, Minneapolis, MH) was used as in receive/transmit mode. Rats were fixed into a custom-built head holder using a mouth bar and ear pins. The total scan time/animal was typically 80–90 min.

Absolute T2 images were acquired by combining Hahn spin-echo data obtained at four TEs between 20 and 110 ms, using a TR of 1.5 s, FOV of 35 mm, matrix size of 128 x 64, slice thickness of 1.5 mm, and an adiabatic BIR-4-refocusing pulse (23) . Tumor volumes were determined from T2-weighted multislice spin-echo images (TR = 2.5 s, TE = 55 ms, matrix size = 256 x 128, FOV = 35 mm, two scans/line, and a contiguous slices of 1-mm thickness; Ref. 20 ). Absolute diffusion images were obtained using a spin-echo sequence (TR = 1.5 s, TE = 55 ms) with four bipolar gradient pairs in each direction. This achieves weighting by the trace of the diffusion tensor (Dav = 1/3 Trace ) in a single acquisition (24) . Data from three acquisitions with different diffusion weighting (b values between 70 and 1420 s/mm2) were used to calculate the absolute Dav images.

A STEAM pulse sequence [TR = 2 s, middle period delay of 30 ms, TE = 2 ms, spectral width (SW = 2.5 kHz), 2048 data points] incorporating an outer volume saturation, a VAPOR water-suppression scheme, and asymmetric excitation pulses (25) was used for determination of tumor lipid concentrations. A voxel was placed within the solid tumor according to the multislice T2-weighted localizer image (TR = 2.5 s, TE = 90 ms, 128 x 128, FOV = 35 mm), and the magnetic field was shimmed using the FASTMAP routine (26) . ADCs for the lipid peaks were quantified by incorporating diffusion sensitizing gradients in the STEAM sequence (TR = 2.5 s, TE = 45 ms) using diffusion time of 76 ms and b values from 0 to 21,316 s/mm2. Lipid concentrations were quantified from STEAM spectra using nonsuppressed water as a concentration reference (0.787 kg/kg fresh tumor tissue; Ref. 27 ). A LASER spin echo sequence (28) was used for semiquantitative assessment of tumor lipids in vivo by 1H NMR spectroscopy (TR = 8.8 s, TE = 32 ms, SW = 2.5 kHz, and 13,000 data points).

For NMR analysis ex vivo, tumor tissue samples (weighing 5–10 mg) were placed into zirconium oxide MAS rotor alongside 10 µl of D2O (deuterium lock reference) containing 10 mM 3-trimethylsilyl propionic acid (chemical shift reference). HRMAS 1H NMR spectra were acquired using a Bruker 600-MHz (Bruker Avance; Bruker GmBH, Rheinestetten, Germany) at 277 K using a conventional solvent suppressed pulse/acquire sequence (based on the start of the two-dimensional NOESY pulse sequence, TR = 2 s, SW = 10 kHz, 32,768 data points, 5-kHz spinning rate; water suppression during the mixing time of 150 ms and relaxation delay). To examine T2 behavior of the lipids, spectra were also acquired with CPMG pulse sequences using an interpulse delay of 500 µs and effective TEs ranging from 10 to 320 ms. In some experiments, the spinning rate and/or temperature were changed as indicated to investigate their effects on spectral characteristics. Free induction decays were multiplied by an exponential apodization function equivalent to 1-Hz line broadening and Fourier transformed to a real transform size of 16,384 data points, giving 8,192 data points of zero filling.

For diffusion spectroscopy, a stimulated echo pulse sequence incorporating bipolar gradients as described by Wu et al. (29) was used with 32 increments of a 53.1 gauss/cm field gradient placed along the magic angle axis. Sine-shaped gradients of 2.5 ms ({delta}/2 for a bipolar sequence) were used with 100 ms of intergradient delay yielding {Delta} of 98.3 ms. The 100-ms delay was repeated three times for each tumor. Thirty-two transients were acquired for each increment using a 16k time domain over a spectral width of 8.4 kHz. ADCs were computed by fitting the NMR signal as a function of b value into a single exponential in XWINNMR (Bruker GmBH).

To additionally characterize the lipid moieties detected in vivo and ex vivo, tissue lipid extracts were reconstituted in 500 µl of CDCl3/CD3OH (3:1) and examined by high resolution 1H and 13C NMR spectroscopy using a triple gradient axis inverse geometry probe (TXI; Bruker GmBH). Solvent suppressed TOCSY and HMBC spectroscopy were performed on these extracts using gradient pulse sequences as described by Braun et al. (30) . For the spectra reported in the results section, a total spin lock time of 80 ms was used alongside sine-shaped gradients of duration 2 ms for the TOCSY (31 , 32) and a 1J(C,H) filter of ~140 Hz and selection for 2J(C, H) and 3J(C, H) ~8 Hz couplings for the HMBC (33) .

In vitro and ex vivo spectra were analyzed in the frequency domain using XWINNMR and in vivo spectra in the time domain using jMRUI software.4

Pattern Recognition of ex Vivo Data.
Ex vivo spectra were integrated across 0.04 ppm of spectral regions between 0.4 and 9.4 ppm using the AMIX software package (Bruker GmBH). The output vector representing each spectrum was normalized across the integral regions, excluding the water resonance, effectively standardizing each integral region to a relative scale of total metabolite concentration, thus avoiding any problems associated with incomplete sample localization in the MAS rotor or variable levels of D2O remaining in the sensitive region. The data sets were imported into the SIMCA package (Umetrics, Umea, Sweden) and then preprocessed using Paretto scaling, where the variance of each integral region was scaled to (1/sk)1/2 so that both high and low concentration metabolites contributed to the partial remission (34) . Each integral region represented an X-variable in the prediction to latent structures through PLSs model, whereas treatment day was represented as a Y vector. This partial remission technique is a regression extension of principal component analysis, identifying variables in which the variation is significantly correlated with whatever the regression is performed against. The goodness of fit algorithm was used to determine whether a correlation was significant (Q2 > 0.097) and, in turn, was calculated by iteratively predicting the Y response for different observations. This algorithm was also used to determine how linear the PLS model was with models possessing a higher Q2 fitting a linear regression model better than those with a low Q2. Metabolic perturbations related to temporal changes and, hence by inference the progression of PCD, were determined from loadings plots. Spectral regions having a modulus loading score of >50% of the maximum loading value for that model were identified as having the most significant changes. Models were built for solvent suppressed CPMG and diffusion weighted spectra.

Histology.
A satellite group of rats were sacrificed by CO2 and transcardial perfusion with PBS for 10 min (30 ml/min) followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) for 10 min (30 ml/min). The fixed brains were removed from the skull, rinsed in PBS, and embedded in ornithine carbamyl transferase medium (Miles, Elkhart, IN) for cryosectioning. Nissl staining was used to reveal cell damage in the sections. The adjoining sections were stained for TUNEL (ApopTag Plus; Oncor, Gaithersburg, MD) to reveal apoptotic nuclei, using a methyl green counterstain. Apoptotic nuclei were counted in tumor tissue from arbitrarily chosen high power fields (x20, Olympus AX-70 microscope; Olympus, Tokyo, Japan).


   RESULTS
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
During MRI of ganciclovir-treated animals, both T2 and Dav images showed patches of eradicating tumor tissue from day 4, with by day 8-treated tumors showing uniformly elevated T2 (Fig. 1, A–C)Citation and Dav (data not shown). Histology of the tumors also confirmed substantial loss of cells, beginning on day 4 of treatment (Fig. 1, D–F)Citation , thus, confirming that the MRI changes above were attributable to cell death. The number of TUNEL-positive cells (Fig. 1, G–I)Citation per high power field increased from 4.0 ± 1.0 by >5-fold by day 6, and tumor volume started to shrink after day 4 of ganciclovir treatment being some 50% less by day 8 (data not shown). These observations confirmed the ongoing PCD (20) .



View larger version (109K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 1. Typical T2-MRI (A–C), histology (D–F) and TUNEL-stained sections from BT4C gliomas. Images are from day 0 (A, D, G), day 4 (B, E, H) and day 8 (C, F, I) of ganciclovir-treated animals. Note that a surface coil was used to acquire MRIs leading to signal drop in the lower brain structures, yet the B1 field obtained by the coil covered the entire tumor volume (A–C). Scale bars in (D–F) are 1 mm and in (G–I) 0.1 mm.

 
1H NMR spectra obtained in vivo from glioma during ganciclovir treatment showed well-resolved peaks at 5.3 and 2.8 ppm, assigned as to CH=CH and CH=CHCH2CH=CH lipid moieties, with these resonances increasing relative to that at 1.3 ppm during treatment (Fig. 2)Citation . The evolution of concentrations of all 1H NMR lipid moieties in vivo is shown in Table 1Citation . It is evident that both 2.8 and 5.3 ppm peaks increased 3-fold by day 8 of ganciclovir treatment, whereas the 0.9 and 1.3 ppm peaks increased by up to 2-fold. Interestingly, the molar increases in these lipid resonances suggest that some 70% of the 1H NMR visible lipid accumulation during PCD was attributable to PUFA build-up. The LASER method was used in conjunction with STEAM method because of sharp volume profiles attained by the adiabatic pulses (28) , and the better signal yield than by STEAM. The LASER spectra from a tumor at treatment day 8 are shown (Fig. 3)Citation both from large and small voxels, the latter covering mainly the T2-hyperintense core. Similar results were obtained from all four day 8 animals scanned. This T2-hyperintense volume has been shown to correspond to cell debris and scar tissue by histology (Fig. 1FCitation ; Ref. 27 ). Interestingly, the spectra are very similar, suggesting uniform presence of 1H NMR lipids in eradicating and dead tumor tissue as well as tissue undergoing PCD. ADC for lipid peaks at 5.3 and 1.3 ppm in vivo did not change as a function of treatment (Table 1)Citation .



View larger version (21K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 2. In vivo 1H NMR spectra from rat gliomas. STEAM 1H NMR spectra were acquired from a tumor-bearing rat before (day 0) and after initiation of ganciclovir treatment as indicated. Peaks are assigned as follows: 1, CH = CH; 2, choline-containing compounds; 3, CH = CHCH2CH = CH; 4, CH2CH = CH; 5, CH2CH2CH2; and 6, CH2CH3.

 

View this table:
[in this window]
[in a new window]

  		Table 1 Concentrations and ADCs of lipid moieties in BT4C gliomas

Concentrations of lipids (in mmol/kg tissue water) and ADCs of 5.3 and 1.3 ppm resonances were determined in vivo as described in "Materials and Methods."

 


View larger version (34K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 3. T2-MRI and in vivo 1H NMR spectra from a glioma. T2-MR images and LASER 1H NMR spectra were acquired from a tumor 8 days after initiation of ganciclovir treatment as described under "Materials and Methods." In A, a spectrum from the entire volume (128 averages) and in B from the T2-hyperintense core (256 averages). Peaks are assigned as in Fig. 2Citation .

 
Ex vivo MAS NMR confirmed the trends detected in the PUFA lipid peaks seen in vivo (Fig. 4)Citation . Altered spin rate (1–9 kHz), temperature variation (277–293 K), or magnetic field strength (9.4 or 14.1 T) did not affect line widths of lipid peaks, indicating that the NMR detectability of lipids, even at 9-kHz spinning speeds, is not affected by weak dipolar effects and/or background gradients and, thus, cannot be from membrane-bound lipid domains showing that these lipids possessed a high degree of rotational freedom. It should be noted that peaks from small molecular weight metabolites were strongly affected by spinning rate (Fig. 5, ACitation versus H). T2 relaxation times did vary for the lipid resonances at 14.1, with olefinic resonances having markedly shorter T2 relaxation time (T2 CH=CH = 37 ms) than the other lipid protons (T2 CH2CH=CH = 150 ms; T2 CH2CH2CH2 = 141 ms; T2 CH3 = 132 ms). Ex vivo diffusion spectroscopy yielded ADCs for lipid resonance (ADCCH=CH = 3.3 ± 1.0, ADCCH2CH=CH = 5.4 ± 0.6; ADCCH2CH2CH2 = 2.5 ± 0.3; ADCCH3 = 3.1 ± 0.5 x 10-11 m2s-1) averaged over the treatment time (n = 10, mean ± SD). Assignments of these PUFA lipid resonances were confirmed by a short-range TOCSY spectrum of intact tumors showing strong cross peaks at 2.8;5.4 ppm caused by olefinic proton resonances (Fig. 6A)Citation . Cross peaks from choline lipids at 4.3;3.7 ppm and multiple cross peaks from myoinositol were also prominent in TOCSY.



View larger version (30K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 4. HRMAS 1H NMR spectra from tumor samples as a function of GVC treatment. HRMAS 1H NMR spectra were acquired at 277 K with v = 5 kHz from normal parietal cortex (Cortex) and tumor samples from days 0 to 8 of ganciclovir treatment. Some of the resonances are assigned as indicated. PtdCholine, phosphatidylcholine; PCholine, phosphocholine.

 


View larger version (24K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 5. Effects of spin rate, temperature and field strength on HRMAS 1H NMR spectra. Tumor sample from day 9 of treatment was studied in a MAS rotor at 14.1 (A–C, F–I) and 9.4T (D). Temperature of the probe and the spin rate (v) are indicated in each panel. * indicates the spinning sidebands visible within the spectral region in A. It should be noted that temperature control is not possible below 286 K at v = 1 kHz in the set-up used.

 


View larger version (32K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 6. A short range TOCSY spectrum from a tumor sample (A) and a HMBC spectrum from tumor lipid extract (B). A, HRMAS TOCSY was acquired from a tumor sample at day 2 of ganciclovir treatment as described in "Materials and Methods." Cross peaks are assigned as indicated. B, HMBC spectrum was collected from pooled tumor lipid extracts from treatment days 2–10 as described in "Materials and Methods." Cross peaks were assigned according to Wilker and Leibfritz (46) . Ala, alanine; CHO, aldehyde; Chol, cholesterol; CholE, cholesterolesters; GSH, reduced glutathione; lac, lactate; Lys, lysine; ml, myo-inositol; Pro, proline; Tau, taurine; TMS, tetramethyl silane.

 
To examine the ability of each lipid resonance to monitor the progression of PCD (12) , PLS-based regression model against time was applied to the ex vivo spectra. The model demonstrated an increase in CH=CH, CH=CHCH2CH=CH, CH2CH2CH2 and CH2CH3 resonances against time (Fig. 7A)Citation . Importantly, CH=CH and CH=CHCH2CH=CH lipid resonances contributed relatively more to the PLS regression loadings vector than the relative resonance intensity when compared with CH2CH2CH2 and CH2CH3, indicating that PCD was correlated with an increase in PUFAs and not just an increase in general lipid moieties (Fig. 7B)Citation . The scalings used in the pattern recognition are most sensitive to the total amount of variation measured in a given spectral region and given the greater intensity of the CH2CH2CH2 lipid resonance compared with the CH=CH and CH=CHCH2CH=CH correlated variation relating to this former resonance is expected to be greater in absolute terms than those affecting the polyunsaturated resonances. However, when comparing these changes in relative terms, whereas there was an increase in the CH2CH2CH2 lipid intensity across the time period, the relative increase in polyunsaturated groups is greater as signified by the greater relative ratio of the loadings for CH=CH/CH2CH2CH2 and CH=CHCH2CH=CH/CH2CH2CH2 compared with a similar analysis on the direct spectra. Models were built also for CPMG spectra using the projections representing either a total echo time of 40 ms yielding similar results (data not shown). This observation additionally corroborates that the lipids were rotationally mobile and existed in a relatively unrestricted environment. Building models using diffusion weighted spectra with a field gradient strength of either 1.1 or 51.3 G/cm, the best model was built, as determined by the goodness of fit, with the smaller field gradient indicating that these lipids were also relatively translationally mobile (data not shown).



View larger version (29K):
[in this window]
[in a new window]
[Download PPT slide]
  		Fig. 7. PLSs regression of spectral profiles obtained form HRMAS 1H NMR spectra against treatment time. In A, the PLS trend identified across the spectral dataset [t(1)] is scaled against a separate vector representing time. This results in a nearly linear trend across the entire dataset. This trend was caused by resonances associated with CH=CH and CH=CHCH2CH=CH as signified in the loading plot in B. Water frequency is indicated by a vertical dotted line in B.

 
To additionally characterize the PUFA lipids lipid extracts of the tumor tissue were examined with one-dimensional and two-dimensional TOCSY, COSY, HSQC, and HMBC NMR methods. The one-dimensional spectra of these extracts demonstrated a 2-fold decrease in total PUFA resonance intensity relative to the CH2CH2CH2 resonance at 1.3 ppm when compared with spectra obtained in vivo and ex vivo. This was also accompanied by increased resonance intensities for acyl and cholesterol side-chain resonances in the lipid extracts, indicating the lipid resonances detected in vivo and ex vivo represented a chemically distinct group from the total lipid content of the tumors. With the HMBC optimized to detect three and four bond couplings between 13C and 1H nuclei, four different CH=CHCH2CH2 lipid groups were detected (Fig. 6B)Citation . The strongest cross peaks for 2.8 and 5.3 ppm 1H resonances arose from 18:1 and 18:2 unsaturated fatty acids with smaller contributions from cholesterol lipids. However, significant quantities of cholesterol were not detected in the tumors in vivo and ex vivo.


   DISCUSSION
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The present data emphasize the need for (a) quantification of lipids by NMR for detection of the cell death in vivo and (b) polyunsaturated nature of lipids accumulating into glioma during ganciclovir-thymidine kinase gene therapy-induced PCD; these lipids being readily NMR observable. Previous data (1, 2, 3 , 9 , 35) have shown that cytoplasmic lipid bodies substantially increase in cells destined to die, and the lack of spectral resolution enhancement of lipid peaks by MAS reported in this study strongly support the claim that the lipid in cytoplasmic vesicles are the source of 1H NMR-detectable resonances. The present novel observations show that the 1H NMR visible PUFAs largely result from 18:1 and 18:2 unsaturated fatty acids.

Progression of PCD is characterized by increased 1H NMR lipid peaks, in particular in PUFAs at 5.3 and 2.8 ppm. The consistent results obtained using the in vivo and ex vivo 1H NMR argue that the lipids detected are in rotationally and translationally unrestricted environment, with high degree of NMR visibility without resolution enhancement by MAS at modest spin rates. Furthermore, these lipids are still readily detectable in CPMG spectra. These spin rates have been shown to cause neither functional (36) nor structural (17) damage for the biological specimens. However, MAS did result in major resolution enhancement of low molecular weight metabolites, particularly myo-inositol and choline-containing compounds. To obtain resolution enhancement for lipids associated with cell membranes, one requires spin rates of an order of magnitude greater than those used here as well as large radio-frequency power. All this is in line with the idea that 1H NMR-detectable lipids are in cytoplasmic lipid vesicles of cells (2 , 9 , 35) and tissue in vivo (3 , 12) , the size of which has been determined to be 5–10 µm in C6 gliomas (3) by diffusion NMR. In these droplets, dipolar couplings and magnetic susceptibility effects are still large enough to preclude observation of fine structure in the lipid resonances, as in the low molecular weight metabolites. Thus, in PCD cytoplasmic lipid body formation recently proposed (37 , 38) is a highly attractive explanation for the 1H NMR spectral changes.

Triglycerides (9 , 18 , 39) and/or other neutral lipids (35) are the predominant lipid groups contributing to the 1H NMR visible lipids detected in cancerous tissue. Accumulation of triglycerides has been attributed to the increase in 1.3 ppm 1H NMR lipid peak in Jurkat T cells (9) and human breast cancer cells during drug-induced cell death (35) . It is well established that triglycerides form the main lipid class in the cytoplasmic lipid vesicles with minor contribution by other classes (37) . The biochemical assays of BT4C gliomas (12) as well as human malignant gliomas (11) have shown that they contain high concentrations of tricycerides, which is virtually absent in the normal brain. During PCD, triglycerides, but also other lipid classes, increase, including cholesterol esters and free fatty acids (12) . The present in vivo and ex vivo results indicate that changes in PUFAs are the most characteristic for PCD in glioma, and in the extracts, this phenomenon is diluted by the extraction of membrane bound lipids containing high portion of saturated lipids. Because only the lipids present in the vesicles are 1H NMR visible in situ (2 , 3 , 9 , 12 , 35) , these PUFA resonances are expected to reside in the cytoplasmic vesicles. This claim is strongly supported by diffusion data both in vivo and ex vivo.

There are two extreme possibilities for the causes of observed change in 1H NMR lipid profile accompanied with an increase in number of cytoplasmic lipid vesicles (12) . These are (a) increase in triglyceride synthesis and (b) change in partitioning of lipids between membranes and cytoplasm or free solution. These two possibilities are not completely mutually exclusive as the release of free fatty acids, especially PUFAs, influence triglyceride metabolism (40 , 41) . Cells submitted to apoptosis by exposure to cytotoxic drugs show inhibition of CDP-choline:1,2-diacyl glycerol choline phosphotransferase leading accumulation of CDP-choline and lipid triglycerides (42 , 43) . This biochemical mechanism would provide increased triglycerides for lipid vesicles and also prevent clearage of cytoplasmic lipids as phosphatidylcholine and subsequent repartition of this compound into cellular membranes. Lipids in the vesicles have greater degree of mobility than in membranous environment, as supported by diffusion and T2 data and, thus, would become NMR observable.

Recent evidence from human breast cancer cells (35) argue for a hypothesis that damaged mitochondrial membranes are the source for NMR observable lipids. In these cells, increase in 1H NMR lipids resonating at 5.35, 1.3, and 0.9 ppm during cytotoxic drug treatment was associated with mitochondrial damage, lipid droplet development, and formation of autophagic vacuoles. Our previous biochemical data from BT4C glioma indicate that phospholipase A2 activity increase during PCD (12) also supporting the claim that membrane lipid partitioning could contribute to the increase in 1H NMR-detected lipids. Interestingly, mitochondrial membranes are known to be rich in PUFAs, with these classes of lipids provide membrane fluidity. During diacylglycerol synthesis, the conversion of 1-acyl glycerolphosphate to 1,2-diacyl glycerylphosphate acyl-CoA containing PUFAs is favored (44) . Mitochondria are also rich in cardiolipin, a phosphatidylglycerol, and significant perturbations of the mitochondrial membrane may accompany the loss of cytochrome c from the membrane during the early stages of PCD. The normal intracellular concentration of K+ inhibits formation of the apoptosome, necessitating extensive and rapid loss of cytochrome c from the mitochondrial membrane for PCD (45) and may provide a mechanism for the detection of cytosolic lipid deposits rich in PUFAs. Thus, the 18:1 and 18:2 fatty acids detected are most likely to be contained in the lipid moieties released from these membranes.

The value of 1H NMR visible lipids in the assessment of cell death pathway is currently debated. The original observation by Blankenberg et al. (7) argued for apoptosis-dependent pathway in the cells showing accumulation of CH2CH2CH2 moieties. More recently, Delikatny et al. (35) reported that in human breast carcinoma cells, these lipids accumulate also during necrotic cell damage and in excised brain tumors saturated lipids as detected by 1H NMR are associated with necrotic build up of histopathology (10 , 11) . In this study, the appearance of PUFAs is evidently associated with progression of PCD. However, in the localized 1H NMR of tumor scar tissue with no signs of viable cells (27) , the PUFA and other lipid resonances are still visible with the same ratios as in late-stage PCD, suggesting that these lipid bodies remain after cell death. Thus, it is not unexpected that the studies using excised tissue arrive to the conclusion above because the separation of necrosis from late-stage PCD by histology may not be feasible. Nevertheless, our results demonstrate a time-window in the cell death when 1H NMR visible PUFA resonances are dynamic.

To conclude, PCD in rat glioma, induced by a robust gene therapy procedure, is associated with PUFA accumulation showing dominant contribution from 18:1 and 18:2 fatty acids. Partitioning of the membrane lipids is proposed as a working hypothesis underlying the observed dynamic 1H NMR changes.


   ACKNOWLEDGMENTS
 
We thank Drs. Michael Garwood, Ivan Tkac, and Rolf Gruetter for providing us with the original versions of the single-voxel 1H NMR pulse sequences and the FASTMAP routine.


   FOOTNOTES
 
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1 Supported by the grants from the Royal Society, the Finnish Cancer Foundation, the Leiras Foundation, and the Academy of Finland. Back

2 To whom requests for reprints should be addressed, at School of Biological Sciences, University of Manchester, 1.33 Stopford Building, Oxford Road, Manchester M13 9PT, United Kingdom. Phone: 44-161-275-1514; Fax: 44-161-275-5363; E-mail: Risto.Kauppinen{at}man.ac.uk Back

3 The abbreviations used are: BIR, B1 insensitive refocusing; MAS, magic angle spinning; NMR, nuclear magnetic resonance; PCD, programmed cell death; PUFA, polyunsaturated fatty acid; HRMAS, high-resolution magic angle spinning; ADC, apparent diffusion coefficient; PLS, partial least square; STEAM, stimulated echo acquisition mode; CPMG, Carr, Purcell, Meiboom, and Gill; LASER, localization by adiabatic selective refocusing; MRI, magnetic resonance imaging; TE, echo time; TR, repetition time; FOV, field of view; TOCSY, total correlation spectroscopy; HMBC, heteronuclear multiple bond correlation; HSQC, heteronuclear single quantum coherence; TUNEL, terminal deoxynucleotidyl transferase-mediated nick end labeling. Back

4 Internet address: http://carbon.uab.es/mrui. Back

Received 12/13/02. Accepted 4/ 7/03.


   REFERENCES
Top
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Hakumäki J. M., Poptani H., Puumalainen A-M., Loimas S., Paljärvi L. A., Ylä-Herttuala S., Kauppinen R. A. Quantitative 1H NMR diffusion spectroscopy of BT4C rat glioma during thymidine kinase-mediated gene therapy in vivo: identification of apoptotic response. Cancer Res., 58: 3791-3799, 1998.[Abstract/Free Full Text]
  2. Barba I., Cabanas M. E., Arus C. The relationship between nuclear magnetic resonance-visible lipids, lipid droplets, and cell proliferation in cultured C6 cells. Cancer Res., 59: 1861-1868, 1999.[Abstract/Free Full Text]
  3. Lahrech H., Zoula S., Farion R., Remy C., Decorps M. In vivo measurement of the size of lipid droplets in an intracerebral glioma in the rat. Magn. Reson. Med., 45: 409-414, 2001.[Medline]
  4. Mountford C. E., Wright L. C. Organization of lipids in the plasma membranes of malignant and stimulated cells: a new model. Trends Biochem. Sci., 13: 172-177, 1988.[Medline]
  5. Ferretti A., Knijn A., Iorio E., Pulciani S., Giambenedetti M., Molinari A., Meschini S., Stringaro A., Calcabrini A., Freitas I., Strom R., Arancia G., Podo F. Biophysical and structural characterization of 1H-NMR-detectable mobile lipid domains in NIH-3T3 fibroblasts. Biochim. Biophys. Acta, 1438: 329-348, 1999.[Medline]
  6. Kauppinen R. A. Monitoring cytotoxic tumour treatment response by diffusion magnetic resonance imaging and proton spectroscopy. NMR Biomed., 15: 6-17, 2002.[Medline]
  7. Blankenberg F. G., Storrs R. W., Naumovski L., Goralski T., Spielman D. Detection of apoptotic cell death by proton nuclear magnetic resonance spectroscopy. Blood, 87: 1951-1956, 1996.[Abstract/Free Full Text]
  8. Blankenberg F. G., Katsikis P. D., Storrs R. W., Beaulieu C., Spielman D., Chen J. Y., Naumovski L., Tait J. F. Quantitative analysis of apoptotic cell death using proton nuclear magnetic resonance spectroscopy. Blood, 89: 3778-3786, 1997.[Abstract/Free Full Text]
  9. Al-Saffar M. M. S., Titley J. C., Robertson D., Clarke P. A., Jackson L. E., Leach M. O., Ronen S. M. Apoptosis is associated with triacylglycerol accumulation in Jurkat T-cells. Br. J. Cancer, 86: 963-970, 2001.
  10. Kuesel A. C., Sutherland G. R., Halliday W., Smith I. C. P. 1H MRS of high grade astrocytomas: mobile lipid accumulation in necrotic tissue. NMR Biomed., 7: 149-155, 1994.[Medline]
  11. Tugnoli V., Tosi M. R., Tinti A., Trinchero A., Bottura G., Fini G. Characterization of lipids from human brain tissues by multinuclear magnetic resonance spectroscopy. Biopolymers, 62: 297-306, 2001.[Medline]
  12. Hakumäki J. M., Poptani H., Sandmair A-M., Ylä-Herttuala S., Kauppinen R. A. 1H MRS detects polyunsaturated fatty acid accumulation during gene therapy of glioma: implications for the in vivo detection of apoptosis. Nat. Med., 5: 1323-1327, 1999.[Medline]
  13. Andrew E. R., Bradbury A., Eades R. G. Removal of dipolar broadening of nuclear magnetic resonance spectra of solids by specimen rotation. Nature (Lond.), 183: 1802-1803, 1959.
  14. Cheng L. L., Lean C. L., Bogdanova A., Wright S. C., Jr., Ackerman J. L., Brady T. J., Garrido L. Enhanced resolution of proton NMR spectra of malignant lymph nodes using magic-angle spinning. Magn. Reson. Med., 36: 653-658, 1996.[Medline]
  15. Millis K., Weybright P., Campbell N., Fletcher J. A., Fletcher C. D., Cory D. G., Singer S. Classification of human liposarcoma and lipoma using ex vivo proton NMR spectroscopy. Magn. Reson. Med., 41: 257-267, 1999.[Medline]
  16. Tomlins A., Foxall P. J. D., Lindon J. C., Lynch M. J., Spaul M., Everett J., Nicholson J. K. High resolution magic angle spinning 1H nuclear magnetic resonance analysis of intact prostatic hyperplasic and tumour tissues. Anal. Comm., 35: 113-115, 1998.
  17. Cheng L. L., Ma M. J., Becerra L., Ptak T., Tracey I., Lackner A., Gonzalez R. G. Quantitative neuropathology by high resolution magic angle spinning proton magnetic resonance spectroscopy. Proc. Natl. Acad. Sci. USA, 94: 6408-6413, 1997.[Abstract/Free Full Text]
  18. Millis K. K., Maas W. E., Cory D. G., Singer S. Gradient, high-resolution, magic-angle spinning nuclear magnetic resonance spectroscopy of human adipocyte tissue. Magn. Reson. Med., 38: 399-403, 1997.[Medline]
  19. Siminovitch D. J., Ruocco M. J., Olejniczak E. T., Das Gupta S. K., Griffin R. G. Anisotropic 2H-nuclear magnetic resonance spin-lattice relaxation in cerebroside- and phospholipid-cholesterol bilayer membranes. Biophys. J., 54: 373-381, 1988.[Medline]
  20. Poptani H., Puumalainen A-M., Gröhn O. H. J., Loimas S., Kainulainen R., Ylä-Herttuala S., Kauppinen R. A. Monitoring thymidine kinase and ganciclovir-induced changes in rat malignant glioma in vivo by nuclear magnetic resonance imaging. Cancer Gene Ther., 5: 101-109, 1998.[Medline]
  21. Pontén U., Ratcheson R. A., Salford L. G., Siesjö B. K. Optimal freezing conditions for cerebral metabolites in rats. J. Neurochem., 21: 1127-1138, 1973.[Medline]
  22. Folch J., Lees M., Sloane-Stanley G. H. A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem., 226: 497-509, 1957.[Free Full Text]
  23. Garwood M., Ke Y. Symmetric pulses to induce arbitrary flip angles with compensation for RF in homogeneity and resonance offsets. J. Magn. Reson., 94: 511-525, 1991.
  24. Mori S., van Zijl P. C. M. Diffusion weighting by the trace of the diffusion tensor within a single scan. Magn. Reson. Med., 33: 41-52, 1995.[Medline]
  25. Tkac I., Starcuk Z., Choi I. Y., Gruetter R. In vivo 1H NMR spectroscopy of rat brain at 1 ms echo time. Magn. Reson. Med., 41: 649-656, 1999.[Medline]
  26. Gruetter R. Automatic, localized in vivo adjustment of all first- and second-order shim coils. Magn. Reson. Med., 29: 804-811, 1993.[Medline]
  27. Hakumaki J. M., Grohn O. H., Tyynela K., Valonen P., Yla-Herttuala S., Kauppinen R. A. Early gene therapy-induced apoptotic response in BT4C gliomas by magnetic resonance relaxation contrast T1 in the rotating frame. Cancer Gene Ther., 9: 338-345, 2002.[Medline]
  28. Garwood M., DelaBarre L. The return of the frequency sweep: designing adiabatic pulses for contemporary NMR. J. Magn. Reson., 153: 155-177, 2001.[Medline]
  29. Wu D., Chen A., Johnson C. S. An improved diffusion-ordered spectroscopy experiment incorporating bipolar-gradient pulses. J. Magn. Reson. (Ser. A), 115: 260-264, 1995.
  30. Braun S., Kalinowski H-O., Berber S. . 100 and more basic NMR experiments. A practical course, VCH Verlagsgesellschaft mBH Weinheim, Germany 1996.
  31. Bax A., Davis D. G. Practical aspects of two-dimensional transverse NOE spectroscopy. J. Magn. Reson., 63: 207-213, 1985.
  32. Hurd R. E. Gradient-enhanced spectroscopy. J. Magn. Reson., 87: 442-428, 1990.
  33. Wilker W., Leibfritz D., Kerssebaum R., Bermel W. Gradient selection in inverse heteronuclear correlation spectroscopy. Magn. Reson. Chem., 31: 287-292, 1993.
  34. Eriksson L., Johansson E., Kettaneh-Wold N., Wold S. . Introduction to multi- and megavariate data analysis USINF projection methods (PCA&PLS), Umetrics Umea, Sweden 1999.
  35. Delikatny E. J., Cooper W. A., Brammah S., Sathasivam N., Rideout D. C. Nuclear magnetic resonance-visible lipids induced by cationic lipophilic chemotherapeutic agents are accompanied by increased lipid droplet formation and damaged mitochondria. Cancer Res., 62: 1394-1400, 2002.[Abstract/Free Full Text]
  36. Griffin J. L., Bollard M., Nicholson J. K., Bhakoo K. Spectral profiles of cultured neuronal and glial cells derived from HRMAS 1H NMR spectroscopy. NMR Biomed., 15: 375-384, 2002.[Medline]
  37. Hakumäki J. M., Kauppinen R. A. 1H NMR visible lipids in the life and death of cells. Trends Biochem. Sci., 25: 357-362, 2000.[Medline]
  38. Murphy D. J. The biogenesis and functions of lipid bodies in animals, plants and microorganisms. Prog. Lipid Res., 40: 325-438, 2001.[Medline]
  39. Rosi A., Luciani A. M., Matarrese P., Arancia G., Viti V., Guidoni L. 1H-MRS lipid signal modulation and morphological and ultrastructural changes related to tumor cell proliferation. Magn. Reson. Med., 42: 248-257, 1999.[Medline]
  40. Clarke S. D. Nonalcoholic steatosis and steatohepatitis. I. Molecular mechanism for polyunsaturated fatty acid regulation of gene transcription. Am. J. Physiol., 281: G865-G869, 2001.
  41. Finstad H. S., Drevon C. A., Kulseth M. A., Synstad A. V., Knudsen E., Kolset S. O. Cell proliferation, apoptosis and accumulation of lipid droplets in U937-1 cells incubated with eicosapentaenoic acid. Biochem. J., 336: 451-459, 1998.
  42. Williams S. N., Anthony M. L., Brindle K. M. Induction of apoptosis in two mammalian cell lines results in increased levels of fructose-1,6-bisphosphate and CDP-choline as determined by 31P MRS. Magn. Reson. Med., 40: 411-420, 1998.[Medline]
  43. Anthony M. L., Zhao M., Brindle K. M. Inhibition of phosphatidylcholine biosynthesis following induction of apoptosis in HL-60 cells. J. Biol. Chem., 274: 19686-19692, 1999.[Abstract/Free Full Text]
  44. Agranoff B. W., Benjamins J. A., Hajra A. K. Lipids Siegel G. J. Arganoff B. W. Albers R. W. Molinoff P. B. eds. . Basic Neurochemistry: Molecular, Cellular, and Medical Aspects, 47-67, Raven Press New York 1999.
  45. Cain K., Langlais C., Sun X. M., Brown D. G., Cohen G. M. Physiological concentrations of K+ inhibit cytochrome c-dependent formation of the apoptosome. J. Biol. Chem., 276: 41985-41990, 2001.[Abstract/Free Full Text]
  46. Wilker W., Leibfritz D. Assignments of mono- and polyunsaturated fatty acids in lipids of tissue and body fluids. Magn. Reson. Chem., 36: S79-S84, 1998.



This article has been cited by other articles:


Home page
 Cancer Res.Home page
C. Denkert, J. Budczies, T. Kind, W. Weichert, P. Tablack, J. Sehouli, S. Niesporek, D. Konsgen, M. Dietel, and O. Fiehn
Mass Spectrometry-Based Metabolic Profiling Reveals Different Metabolite Patterns in Invasive Ovarian Carcinomas and Ovarian Borderline Tumors.
Cancer Res., November 15, 2006; 66(22): 10795 - 10804.
[Abstract] [Full Text] [PDF]

Home page
 Physiol. GenomicsHome page
H. J. Atherton, N. J. Bailey, W. Zhang, J. Taylor, H. Major, J. Shockcor, K. Clarke, and J. L. Griffin
A combined 1H-NMR spectroscopy- and mass spectrometry-based metabolomic study of the PPAR-{alpha} null mutant mouse defines profound systemic changes in metabolism linked to the metabolic syndrome
Physiol Genomics, October 11, 2006; 27(2): 178 - 186.
[Abstract] [Full Text] [PDF]

Home page
 Phil Trans R Soc BHome page
J. L Griffin
The Cinderella story of metabolic profiling: does metabolomics get to go to the functional genomics ball?
Phil Trans R Soc B, January 29, 2006; 361(1465): 147 - 161.
[Abstract] [Full Text] [PDF]

Home page
 FASEB J.Home page
L. G. Astrakas, I. Goljer, S. Yasuhara, K. E. Padfield, Q. Zhang, S. Gopalan, M. N. Mindrinos, G. Dai, Y.-M. Yu, J. A. J. Martyn, et al.
Proton NMR spectroscopy shows lipids accumulate in skeletal muscle in response to burn trauma-induced apoptosis
FASEB J, September 1, 2005; 19(11): 1431 - 1440.
[Abstract] [Full Text] [PDF]

Home page
 Clin. Cancer Res.Home page
L. G. Astrakas, D. Zurakowski, A. A. Tzika, M. K. Zarifi, D. C. Anthony, U. De Girolami, N. J. Tarbell, and P. M. Black
Noninvasive Magnetic Resonance Spectroscopic Imaging Biomarkers to Predict the Clinical Grade of Pediatric Brain Tumors
Clin. Cancer Res., December 15, 2004; 10(24): 8220 - 8228.
[Abstract] [Full Text] [PDF]

Home page
 J. Biol. Chem.Home page
K. K. Lehtimaki, P. K. Valonen, J. L. Griffin, T. H. Vaisanen, O. H. J. Grohn, M. I. Kettunen, J. Vepsalainen, S. Yla-Herttuala, J. Nicholson, and R. A. Kauppinen
Metabolite Changes in BT4C Rat Gliomas Undergoing Ganciclovir-Thymidine Kinase Gene Therapy-induced Programmed Cell Death as Studied by 1H NMR Spectroscopy in Vivo, ex Vivo, and in Vitro
J. Biol. Chem., November 14, 2003; 278(46): 45915 - 45923.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Griffin, J. L.
Right arrow Articles by Kauppinen, R. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Griffin, J. L.
Right arrow Articles by Kauppinen, R. A.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online