TO THE EDITOR:
We read with interest the article by Wallace et al. (1) describing PET imaging of pulmonary fibrosis in a rabbit model. We were pleased to find that, in general, their data confirmed our own results, obtained in a pilot study and reported in abstract form (2). In their article, Wallace et al. raise several important questions, which we believe our pilot study might help resolve. In addition to PET scanning, we have performed cellular resolution autoradiography and some biodistribution studies of both the cis- and the trans-isomer of fluoroproline (FP) and believe that this additional information will help in the interpretation of the PET signal in this potentially important application.
Our basic methodology was similar to that of Wallace et al. (1). A lower dose of microcrystalline silica particles, 50 mg of 5-μm particles in 0.5 mL of normal saline, was instilled into the right upper lobe of anesthetized rabbits’ lungs under direct vision through the biopsy channel of a neonatal pediatric bronchoscope. At intervals after instillation, the rabbits were anesthetized again, and 1 ear vein was cannulated for injection of radioactive marker. The artery in the contralateral ear was cannulated for withdrawal of blood samples. Each rabbit was positioned with its thorax within the field of view of a PET scanner. Several animals were scanned simultaneously.
Cis- or trans-18F-FP was prepared as previously described (3), dissolved in normal saline, and injected into the ear vein of each rabbit. Emitted radioactivity was measured by PET scanning for 6 successive frames of 15 min. Unlike Wallace et al. (1), we had few problems in localizing the 18F signal to the challenged area. 18F-FP data were corrected for attenuation using the transmission scan data, and the ratio of the uptake rate of radioactivity in the challenged (right lobe) regions of interest to that in the unchallenged (left lobe) regions of interest was calculated. Individual animals were scanned up to 6 times over the 13-wk course of the experiment (measured from instillation of silica).
Silica challenge of the lung increased the uptake of trans-18F-FP, first detectable at 13 d. The PET signal peaked at 6–8 wk and by 13 wk, in contrast to the findings of Wallace et al. (1), had declined significantly. However, at that time the signal was still above baseline. ANOVA with least significant difference testing (P < 0.05) showed that 18F-FP uptake at 41 and 54 d was significantly different from that at day 0; that 18F-FP uptake at 54 d was significantly different from that at 0, 5, and 7 d; and that 18FP uptake at 41 d was significantly different from that at 0, 5, 7, and 13 d. In a single study with the cis-isomer, uptake at 6 wk differed little from that observed after the trans-isomer. The difference in time course of the signal between our study and that of Wallace et al. might have been due to the scanning interval (90 min, compared with 180 min). The signal acquired over the shorter interval will reflect primarily uptake, whereas that over the longer interval will reflect protein synthesis (4) (but only with the cis-isomer). However, an alternative explanation is that the maintenance of the signal perceived by Wallace et al. reflects the increased density of scar tissue rather than active collagen synthesis, as no dynamic analysis of their data was presented and the data do not appear to have been corrected for density.
In our pilot study, tissue samples were taken and counted in a γ-well counter after cis- and trans-18F-FP at 6–8 wk after challenge—a single rabbit in each case. In 2 other rabbits, high-performance liquid chromatography analysis was performed on samples of plasma and urine after injection of trans-18F-FP.
After both isomers of 18F-FP, uptake was clearly greater in the challenged lung than in the nonchallenged lung, confirming the results obtained by PET scanning. Little or no uptake occurred in heart or muscle. Nor was there any uptake into bone, as would have occurred should the 18F-label have dissociated from the 18F-FP during the scanning period. In liver and kidney, uptake was markedly greater for the cis-isomer than for the trans-isomer. However, this difference would not have affected the pulmonary signal.
As indicated above, counting of tissue samples showed that there was no appreciable uptake into bone with either isomer, suggesting that defluorination during the study was minimal. This finding was confirmed by high-performance liquid chromatography after trans-18F-FP injection. Only the parent compound could be detected in blood and urine. No free fluoride could be detected.
In a single rabbit 13 wk after instillation of microcrystalline silica, 3H-proline was coinjected with the trans-18F-FP. After the PET scan, the animal was killed and the lungs were immediately removed and inflated to a pressure of 15 cm with formol saline and processed for microautoradiography.
In sections from the control lung, the architecture was normal and the grains developed by autoradiography were distributed randomly, with no evidence of localized uptake. The challenged lung showed disruption of the architecture and massive interstitial thickening. Autoradiographic grains were associated with the cellular component of the lesion, principally fibroblasts (probably myofibroblasts). The acellular fibrotic area in the center of the lesion showed no increase in radiolabeling.
In conclusion, our pilot studies agree with the findings of Wallace et al. (1) that, in response to a fibrotic stimulus to the lung, it is possible to monitor increased uptake of 18F-FP into the challenged region by external imaging using PET. Our preliminary studies show that this signal likely reflects upregulation of proline transport into fibroblasts, rather than collagen synthesis. This possibility needs to be confirmed in further studies. Cis-proline and trans-proline appear to give similar signals up to 90 min. This finding is consistent with the signal’s reflecting proline transport rather than collagen synthesis, as there is evidence that although the isomers behave similarly in the former, they differ appreciably in the latter (4,5). PET imaging of 18F-FP clearly has considerable potential in monitoring the fundamental processes involved in the development of scarring of the lung and other tissues in living animals and perhaps in patients. We look forward to further development of this technique.
REFERENCES
REPLY:
The pilot study of 18F-fluoroproline (FP) imaging of silica instillation-induced pulmonary fibrosis in rabbits (1), supplemented with information in the above letter to the editor and our investigations (2), suggest that PET imaging using 18F-FP has the sensitivity to detect metabolic events involved in pulmonary response to some respirable quartz dust exposures. The studies used different isomers of FP, cis in ours (2) and mostly trans in the other, as reported in the letter, and images were read for different scanning intervals. These factors may be involved in the differences seen in response versus time after dust challenge. All emission data in our study were corrected for tissue attenuation, as reported (2), and planned analyses of shorter scanning-time data acquired in our study may clarify the question, but the much stronger dust doses administered in our study (in which instillation was purposefully not directed to a particular lung or lobe as an added experimental blindfold) might be a significant factor.
The range of conditions of exposure and response under which early or progressive stages of disease such as silicosis or asbestosis could be so detected or evaluated as metabolic events (e.g., exacerbated uptake or incorporation in collagen synthesis of the labeled proline analog) is to be determined. Questions of specificity persist: tritiated-proline autoradiography has been used for some time to study fibroblast synthesis of collagen (3), but little quantitative information exists on the fractional distribution of the label between alveolar cells involved in the inflammatory response (e.g., neutrophils) versus interstitial fibroblast incorporation of the labeled proline. 18F-FDG has been shown to be taken up specifically by inflammation-related neutrophil influx in response to microcrystalline silica challenge (4). Further, it might be important to use tritiated FP itself in addition to tritiated proline in such studies. A report on diastereomeric effects on 4-18F-FP metabolism and uptake in tumors in mice (5) indicated differences between uptake and incorporation for different FP isomers.
The questions of the adequacy of the sensitivity and specificity of the method (e.g., for application in studies of occupational exposure and disease) appear to be amenable to investigation. If found to be accurate, the technique potentially would be highly useful diagnostically. However, as a cautionary note to clinical trials: We have seen evidence of eosinophilic endarteritis in some lung histopathology sections of our rabbits. We are investigating this finding to determine whether the vascular inflammation was endemic to the animals or was caused by FP or by nonradiolabeled contaminants in our test article, such as residue from protective groups on the precursor used in 18F-FP synthesis.