| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | RSS | TABLE OF CONTENTS |
|
|
|
||||||||
|
||||||||||
Basic Science Investigation |
1 Department of Nuclear Medicine, University Hospital Gasthuisberg and Catholic University Leuven, Leuven, Belgium; 2 Molecular Small Animal Imaging Center, University Hospital Gasthuisberg and Catholic University Leuven, Leuven, Belgium; 3 Center of Human Genetics, University Hospital Gasthuisberg and Catholic University Leuven, Leuven, Belgium; 4 Laboratory for Radiopharmacy, University Hospital Gasthuisberg and Catholic University Leuven, Leuven, Belgium; 5 Department of Hematology, University Hospital Gasthuisberg and Catholic University Leuven, Leuven, Belgium; and 6 Department of Pathology, University Hospital Gasthuisberg and Catholic University Leuven, Leuven, Belgium
Correspondence: For correspondence contact: Lieselot Brepoels, MD, Department of Nuclear Medicine, University Hospital Gasthuisberg, Leuven Herestraat 49, 3000 Leuven, Belgium. E-mail: lieselot.brepoels{at}uz.kuleuven.be
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Key Words: therapy response • 18F-FDG • corticosteroids • mice • small-animal PET
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
During the last few years, several clinical studies have demonstrated that 18F-FDG PET significantly improves the accuracy of initial staging, management of recurrent cancer, and monitoring of therapy response of lymphoma (1–4). 18F-FDG PET is a functional imaging technique that reflects the glycolytic rate of tissues and has been used to measure the increased metabolic demand in tumor cells. Currently, the use of PET for response assessment is changing from evaluation at the end of treatment to prediction of tumor response early during the course of therapy. However, early after the initiation of therapy, it is not uncommon to see residual 18F-FDG uptake in the tumor mass, and cutoff values (percentage decrease from baseline or amount of residual uptake) will probably be necessary to differentiate responders from nonresponders. To be a reliable predictor of response, 18F-FDG uptake should reflect changes in the amount of viable tumor cells. However, the rapid decrease in tumor glucose metabolism after effective therapy is only partially concordant with 18F-FDG uptake in viable tumor cells. The inflammatory stromal cells, which appear as a result of chemotherapy or radiotherapy, also show a high intracellular accumulation of 18F-FDG and may even show a higher uptake than that of the viable tumor cells, making it difficult to correctly estimate the viable tumor fraction (5). The aim of this study was to evaluate whether the administration of corticosteroids can suppress this inflammatory reaction and, in turn, improve the correlation of PET response and tumor cell kill.
The experiments were performed using a previously validated mouse model in which severe combined immunodeficiency (SCID) mice were inoculated subcutaneously with Daudi B-cells, a human lymphoma cell line derived from a Burkitt lymphoma (6). All mice were treated with cyclophosphamide, and half of the number of mice also received hydrocortisone over 5 d, analogous to the present standard therapy in humans.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Experimental Design
Seventy-one mice were used for analysis (Table 1). Six mice were sacrificed after small-animal PET, when their tumor diameter had reached 15 mm, without receiving any treatment. Their tumors were dissected for flow cytometry (n = 4) and histopathology (n = 2), and these results were used as reference values (day 0).
|
18F-FDG uptake over time was evaluated by serial small-animal PET (5 mice in each group) on D0 (before treatment), D+6, D+9, D+13, and D+16. After their last scan (D+16), these mice were sacrificed and their tumor was dissected for flow cytometry or histopathology.
Four mice of each group were sacrificed after small-animal PET on D+2, D+4, D+6, D+9, and D+13, and their tumors were dissected for flow cytometry. One mouse in group A and 2 mice in group B were sacrificed on D+2, D+4, D+6, D+9, and D+13, and their tumors were dissected for histopathology.
Small-Animal PET Scanning
After overnight fasting, small-animal PET was performed using a Focus 220 microPET (Siemens Medical Solutions USA, Inc.). The spatial resolution of this PET system is 1.6 mm, and acquired images were reconstructed with ordered-subsets expectation maximization. Simulations with spheres of different diameters using these settings showed that the effects of partial volume on 18F-FDG uptake are minimal in lesions with a diameter larger than 3 mm.
After sedation by gas anesthesia (isoflurane, Forene; Abbott), tumor dimensions were measured (by caliper), and body weight and glycemia were determined. Sixty minutes after injection of 8–11 MBq 18F-FDG via a tail vein, small-animal PET was performed (10 min), on a single bed position with the tumor in the center of the field. All animals received an intramuscular injection of furosemide (40 mg/kg, Lasix; Aventis) in the contralateral thigh at the same time as the tracer injection to reduce reconstruction artifacts caused by the high concentration of 18F-FDG in the urine, and the bladder was carefully emptied before scanning.
Determination of Different 18F-FDG PET Parameters
In the 10 mice that were used for serial follow-up (5 mice in group A and 5 in group B), the mean standardized uptake value (SUVmean), the maximal SUV (SUVmax), and the metabolic tumor volume (Volmetab) were measured at each time point. SUVs were automatically calculated using the formula SUV = measured activity concentration in the tumor (Bq/g) x body weight (g)/injected activity (Bq). Volume contouring was done on a visual basis (3-dimensional isocontour), with in-house software (8). The total lesion glycolysis (TLG) was defined as SUVmean x Volmetab. The change in TLG (
TLG) was calculated and expressed as a percentage of the baseline TLG on D0: TLG = (TLGx – TLG0)/TLG0 x 100, with TLGx being the 18F-FDG uptake on day x (9). The percentage decreases in SUVmean, SUVmax, and Volmetab were calculated in an analogous way. Whereas SUVs represent the density of cells in a certain tumor volume (cell fraction), TLG provides a measurement of the total 18F-FDG uptake in the whole tumor lesion (absolute number of cells). To correct for differences in administered dose (paravenous injections became more frequent after several injections in the same mouse), all SUVs were normalized by dividing SUVs of the tumor by the corresponding SUVmean of a standard region in the liver (10).
In sacrificed mice, the 18F-FDG activity in the tail was measured, decay-corrected, and subtracted from the initially injected dose. SUVs were further calculated in an analogous way. After dissection, the activity in the tumor was counted in a well counter (Wizard; Wallac).
Flow Cytometry
To allow comparison of the different contributing cell fractions with 18F-FDG uptake, mice were sacrificed by decapitation a few minutes after small-animal PET. The tumor was carefully dissected and was teased apart to obtain single cell suspensions suitable for flow cytometry (Becton Dickinson immunocytometry systems).
Cells (1 x 106 cells/sample) were stained with Peridinin–chlorophyll–protein (PerCP)-conjugated antihuman CD45 monoclonal antibody ([mAb] Becton Dickinson) and fluorescein isothiocyanate (FITC)-conjugated antimouse CD45 mAb (Serotec) to calculate the number of cells contributed by the human tumor cells versus the stromal mouse cells. A second sample was stained with annexin-V–FITC and propidium iodide (PI) (Clontech). Necrotic cells are PI positive and annexin-V positive, whereas apoptotic cells are annexin-V positive and PI negative (6). Flow cytometry was performed by a fluorescence-activated cell sorter to acquire the cell fraction composition for each tumor (Cellquest software; Becton Dickinson Bioscience). From further measurements (not shown), we learned that the mouse CD45-positive fraction consisted almost exclusively of viable cells and that with the human CD45 mAb the measured composition consisted of viable and apoptotic and necrotic human cells. Therefore, the percentage of viable tumor cells was calculated as the percentage of human cells minus the sum of apoptotic and necrotic cells.
In some mice (2 nonsteroid- and 2 steroid-treated mice on D+2, D+4, D+6, and D+9), we also performed supplementary experiments for further differentiation of the inflammatory response. By gating on the mouse CD45- positive cells, we further differentiated F4/80-positive cells (macrophages), Ly6 (Gr-1)-positive cells (granulocytes) (11), and NK1.1-positive cells (NK cells).
Histology
After careful dissection, the tumors were fixed and embedded in paraffin. Sections were stained with hematoxylin and eosin, and immunostaining was performed using CD20 mAb, a pan B-cell marker to identify the human tumor cells. The sections were also stained with Ki67 (Mib-1) for identification of proliferating cells.
Tumors were semiquantitatively scored for CD20-positive tumor cells, Ki67-positive cells, apoptotic/necrotic cells, inflammatory cells, and fibrosis (–, almost no cells; +, some cells; ++, lots of cells; and +++, almost all cells).
Statistics
For each time point, the mean ± SEM of the different variables (SUVs, viable tumor cells, stromal host cells, necrotic and apoptotic cells) in both treatment groups were calculated and expressed graphically. A 2-tailed P value of P 
0.05 was considered statistically significant. When a significant difference between the 2 groups was suspected, unpaired Student t tests were performed. Bonferroni correction was subsequently used to reduce falsely significant results when multiple comparisons were suspected, and corrected P levels were reported as Pcorrected. To compare overall reduction in 18F-FDG between both groups over time, the area under the curve (AUC) of the different 18F-FDG PET parameters was calculated and compared with the unpaired Student t test.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
|
0.01). On D+13, the inflammatory cell fraction had returned almost to baseline values and was similar in both groups. No representative cell suspensions for flow cytometry could be obtained on D+16, because there was only a small amount of tumor tissue that could not be differentiated from surrounding fibrosis.
Serial Follow-up by Small-Animal PET Scanning
On D+6, there was a fast decrease in all measured parameters on small-animal PET in both treatment groups, in parallel with the reduction in the viable tumor cell fraction (Figs. 2A and 2B). In group A, the initial decrease in SUVmean and SUVmax was followed by a stabilization and a mild increase on D+9, despite a further reduction in the viable tumor cell fraction. In group B, the initial decline in SUVmean and SUVmax was followed by a further decrease on D+9, however, more limited. There seemed to be a difference at the P 0.05 level in SUVmax on D+9 between both groups (P = 0.032; –48.6% vs. –71.0%) as well as in TLG on D+9 (P = 0.017; –88.0% vs. –94.7%), but this was not significant after Bonferroni correction (Pcorrected 0.0125). There was also a tendency toward a difference in decline in SUVmean and Volmetab (P = 0.053 and P = 0.060). Time–activity curves in group B were significantly lower compared with those in group A (P = 0.023 and P = 0.034 for AUC SUVmax and AUC TLG).
|
Ex vivo evaluation of 18F-FDG uptake, as counted in the well counter, was very similar to the results of small-animal PET (results not shown). There was also no systematic difference in the glucose levels between mice that received steroids compared with mice that received only chemotherapy. Glucose correction did not change the curves of measured PET parameters and resulted in the same conclusions as noncorrected PET results.
Histopathology
In untreated animals, tumors were composed of a massive proliferation of CD20-positive cells with a high proliferation rate (Ki67 highly positive) and the presence of multiple mitotic figures (Table 2; Fig. 3A). Almost no stromal reaction is found, with very few mononuclear cells and virtually no polymorphonuclear cells.
|
|
To confirm these results, we performed supplementary flow cytometry for further differentiation of the stromal reaction (Fig. 4). We counterstained mouse CD45-positive cells with phycoerythrin-conjugated F4/80 (macrophages), Ly6 (Gr-1) (granulocytes), and NK1.1 (NK cells). We observed a change in the proportion of macrophages/granulocytes, with a shift from mainly macrophages in group A to a mainly polymorphonuclear infiltration in group B. Given the small number of samples (only 2 samples per time point), we did not perform a statistical analysis on these data.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Two conclusions may be drawn from small-animal PET measurements. First, the reduction in 18F-FDG uptake measured by the different PET parameters was more important in steroid-treated mice, resulting in a response curve that was systematically located below the curve of the nonsteroid regimen. This may correspond to the known lympholytic effect of steroids, especially in lymphoma, leading to a more pronounced tumor response. Second, SUVs tended to increase on D+9 after chemotherapy, and this effect is suppressed by the administration of steroids, suggesting a lower 18F-FDG uptake caused by the inflammatory influx in steroid-treated mice.
Flow cytometric analyses were performed for evaluation of the different cell fractions that contribute to the observed 18F-FDG uptake. During the first 6 d after chemotherapy, the decrease in SUVs represents very well the decrease in the viable tumor cell fraction, primarily the only cell population at that time. On D+9 after the administration of cyclophosphamide, there was a transient stromal reaction, and SUVs tended to increase, although a further decrease in the viable cell fraction was seen. However, the increase of SUVs was mild, which is probably due to the fact that the loss of viable tumor cells on D+9 outweighs the infiltration of inflammatory cells. These results correspond very well with the results reported by Spaepen et al. (6), validating our technique and our reproducibility.
Administration of steroids for 5 d resulted in a suppression of the peak inflammatory reaction on D+9, corresponding to a suppressed 18F-FDG uptake on that day. However, steroid administration also resulted in an increased stromal reaction at earlier time points (D+6). The stromal reaction as seen after cyclophosphamide administration was changed under the influence of corticosteroids, with a shift from a mononuclear reaction to a predominately granulocytic reaction. This shift can be explained by the properties of steroids to provoke a mobilization and activation of granulocytes, leading to a higher level of circulating granulocytes and a higher and faster penetration into the tumor lesion, as shown by Maruyama et al. (12).
In contrast to the stimulating effect on granulocytes, there is a reported inhibitory effect of steroids on the differentiation of macrophages (13), monocytes (14), lymphocytes (15,16), and NK cells (17,18) and on the access to an inflammatory site. Macrophages and monocytes play an important role, as they are involved in the clearance of cellular debris and apoptotic cells and are, therefore, responsible for a substantial amount of the stromal reaction. The upregulation of granulocytes and the downregulation of mononuclear cells can explain the changing macrophage/granulocyte ratio that was observed on histology and confirmed by flow cytometry. As different types of leukocytes are marked by a different glucose metabolism, a changed composition of the inflammatory influx under steroid administration can result in an altered 18F-FDG uptake (19).
Several authors have shown that 18F-FDG uptake in leukocytes is strongly dependent on their level of activation and on complex interactions between cells (20–22). Granulocyte function is relatively refractory, whereas monocyte/macrophage function is particularly suppressed by corticosteroids (23). Therefore, it is plausible that 18F-FDG uptake in the steroid-treated mice is influenced not only by the amount and type of cells but also by a changed activation state of the leukocytes.
The choice for cyclophosphamide as the cytotoxic treatment was inspired by the administration of this drug in clinical settings, but we are aware that this compound also has an immunosuppressive effect (24). It is possible that this also affects the stromal reaction in our tumor model, but there is no evidence that this effect is different in steroid- and nonsteroid-containing regimens.
Similar to the results of Spaepen et al. (6), we found no clear differences in the apoptotic cell fraction under chemotherapy, and the necrotic cell fraction—though increasing after chemotherapy—showed a rather large variability. The mechanical stress these cells undergo—caused by the mechanical teasing apart of the cells—and the delay between excision of the tumors and flow cytometry probably induced a certain degree of spontaneous apoptosis and necrosis. As the measurements of necrosis and apoptosis are directly reflected in the calculation of the viable tumor cell fraction, this may also explain the absence of a clear difference between the viable cell fraction in the nonsteroid- and the steroid-containing regimen (Fig. 1A), despite the differences in 18F-FDG uptake. Measurement of the inflammatory mouse fraction does not have this limitation; therefore, we based our conclusions primarily on the results of the inflammatory mouse fraction as well as on the histologic sections.
Our tumor model involves an immunodeficient animal (SCID mice) without a functional T- or B-cell compartment. Therefore, it is not possible to evaluate the lymphocytic cell fraction, and our data cannot be generalized to immunocompetent mice or humans. Despite these limitations, these mice have a normal macrophage, granulocyte, and NK function, and an inflammatory reaction was shown in this tumor model in earlier experiments (6). However, immunodeficient animals are inherent to the study of human cell lines in vivo, and no lymphoma model in immunocompetent mice has been described. Our hypothesis is that results in immunocompetent mice will be more dramatic, because of a more important inflammatory response, but will lead to the same conclusions.
Although there is no consensus yet on what the thresholds should be for early therapy response assessment by 18F-FDG PET, guidelines have been proposed already by the European Organization for Research and Treatment of Cancer (EORTC) PET group (25) and were used in clinical trials to define early PET response. According to the EORTC PET group, a decrease in the SUV from baseline of at least 15% after 1 cycle and 25% after several cycles is necessary to obtain a partial response. In this context, a transient influx in inflammatory cells—even if it accounts for a rather mild increase in 18F-FDG uptake—can be of importance, as well as the effect of steroids, which are often part of the treatment regimen in lymphoma patients.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONCONCLUSIONReferences |
|---|
Related articles in JNM:
This article has been cited by other articles:
|
|
 |
|
  L. Brepoels, S. Stroobants, G. Verhoef, T. De Groot, L. Mortelmans, and C. De Wolf-Peeters 18F-FDG and 18F-FLT Uptake Early After Cyclophosphamide and mTOR Inhibition in an Experimental Lymphoma Model J. Nucl. Med., July 1, 2009; 50(7): 1102 - 1109. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | RSS | TABLE OF CONTENTS |
| JOURNAL OF NUCLEAR MEDICINE TECHNOLOGY | THE JOURNAL OF NUCLEAR MEDICINE |
, group A) and mice treated with chemotherapy and hydrocortisone (
, group B) and expressed as mean percentage ± SEM. x-axis: different time points (days after treatment); y-axis: percentage cell fraction out of total, as measured by flow cytometry. *Significant differences after Bonferroni correction between both treatment groups.
