The results of this study indicate that multiphoton FLIM provides structural (cellular-level morphology in three spatial dimensions) and metabolic (NADH lifetime) information that discriminates between normal tissue and pre-cancers in vivo
. The morphological differences identified with pre-cancer development (increased nuclear diameter and epithelial thickness) are consistent with a previous multiphoton microscopy study of the DMBA-treated hamster cheek pouch (28
). These morphological changes are also consistent with previous histopathology studies (16
). However, multiphoton FLIM has the advantage of resolving these structural differences in vivo
without the need for physical sectioning, and providing functional images of tissue metabolism with the NADH fluorescence lifetime. The protein-bound NADH fluorescence lifetime discriminated normal vs. high grade and normal vs. low grade pre-cancers in vivo
(). The amount of protein-bound NADH relative to free NADH (α2
) also decreased with low grade pre-cancer compared to normal ().
Cell culture perturbation studies () revealed that the protein-bound NADH lifetime component (τ2
) increases with the addition of 3-bromopyruvate and decreases with the addition of CoCl2
(consistent with previous work (24
)), and these lifetime changes increase and decrease with the redox ratio, respectively. It is expected that the addition of 3-bromopyruvate causes inhibition of glycolysis, thus resulting in an increase in the redox ratio while COCl2
causes inhibition of oxidative phosphorylation thus decreasing the redox ratio in cell culture. It should be emphasized that the results in are not intended for comparison of the merits of the redox ratio vs. the fluorescence lifetime. The purpose of showing the redox ratios was to use an independent method other than the fluorescence lifetime to show that there are changes in the redox state of the cells with the specific metabolic perturbations used in this study.
Changes in cellular metabolism with cancer development in vivo
are more complex than the simple chemical perturbations to cellular metabolism in cell monolayers. Alterations in cellular metabolism with neoplasia development in vivo
can be due to a number of factors including genetic changes, changes in tissue vascularization, and changes in metabolic demand (6
). The purpose of was to show that the fluorescence lifetime is sensitive to changes in cellular metabolism in simple cell monolayers using known chemical perturbations. However, the cell monolayer results cannot be directly applied to the in vivo
results due to the complex changes in metabolism with neoplasia development in vivo
The short lifetime component (τ1
) and long lifetime component (τ2
) of the normal epithelial cells measured in vivo
in the hamster cheek pouch are within the range of published values of free NADH and protein-bound NADH, respectively, in cell culture, tissue slices and in vivo
). However, the published lifetimes of free and protein-bound NADH vary widely depending on the biological system investigated. It is also possible that flavin adenine dinucleotide (FAD) could contribute to the measured lifetime at a two-photon excitation wavelength of 780 nm. However, in vivo
multiphoton fluorescence images collected at 900 nm excitation (optimal for the two-photon excitation of FAD (26
)) did not show any measurable fluorescence with our in vivo
FLIM system, indicating that FAD is unlikely to contribute to the observed lifetime at 780 nm excitation in the hamster cheek pouch. Note that the FLIM system used for the in vivo
experiments was not optimized for 900 nm excitation (due to low laser power and detection efficiency), but the system used for the cell culture experiments was optimized for collection at both 780 nm and 890 nm, so FAD fluorescence was observed in the cell culture experiments. Another source of fluorescence at this excitation wavelength is NADPH. Previous studies have shown that the concentration of NADH is about 5 times greater than NADPH (40
), the NADH quantum yield is 1.25 – 2.5 times greater than NADPH (41
), and metabolic perturbations produce fluorescence changes dominated by NADH (42
). Thus, its likely that NADPH is responsible for a low, relatively constant fluorescence background in these studies (37
Normal and pre-cancerous tissues were differentiated with the cellular protein- bound NADH lifetime (τ2
), with low grade and high grade pre-cancerous tissues having a lower τ2
than normal (Table 2). The Warburg effect predicts that neoplasias favor glycolysis over oxidative phosphorylation under aerobic conditions (6
), and given the results of the cell perturbation study (), the measured decrease in protein-bound NADH lifetime with neoplasia is consistent with increased levels of glycolysis. This shift to glycolysis could be due to either the Warburg effect or the Pasteur effect (43
), which is increased glycolysis due to hypoxia. Changes in the distribution of NADH enzyme binding sites associated with preferred metabolic pathways in neoplastic tissues (10
) may be responsible for the change in protein-bound NADH lifetime with pre-cancer. However, further studies are required to test this hypothesis.
Increased intracellular τ2 and α2 variability were observed in pre-cancerous tissues, and low grade pre-cancers had increased intracellular α2 variability compared to high grade pre-cancers (). This could be due to heterogeneous NADH binding and microenvironments within the cells. FLIM studies of pre-cancerous cells in vivo at higher magnifications with perturbations to cytosolic and mitochondrial metabolism could further investigate the heterogeneity of lifetimes within individual cells.
The lifetime of low grade pre-cancers was found to decrease with depth in the epithelium, while the lifetime of normal and high grade pre-cancers did not change with depth (). Neoplastic cells originate near the basement membrane and can move progressively upward through the epithelium (9
). In a low grade pre-cancer, the epithelium is only partially occupied by the less differentiated neoplastic cells, while in a high-grade pre-cancer the full thickness of the epithelium is occupied by these cells (44
). It may be speculated that increased heterogeneity in cell populations in low grade pre-cancers may be partly responsible for the variation in lifetime with depth in these tissues.
Multiphoton FLIM is ideal for pre-cancer detection in squamous epithelial tissues because the penetration depth of multiphoton FLIM (dependent on the optical properties of the organ site, but up to 0.3 mm in human skin (45
), for example) is comparable to the thickness of the epithelium in a variety of tissues including the cervix and oral cavity (46
). Multiphoton FLIM could also be used for other clinical applications, including non-invasive glucose monitoring in diabetics (47
) and metabolic monitoring for tumor therapy. In the near-term, the important findings from multiphoton FLIM studies could guide the design and development of practical time-gated fluorescence detection schemes for clinical applications. In long-term, portable technology could be engineered to enable multiphoton FLIM in a clinical setting.