Imaging of hypoxia, oxygen consumption and recovery in vivo during ALA-photodynamic therapy using delayed fluorescence of Protoporphyrin IX

Abstract

Background: Hypoxic lesions often respond poorly to cancer therapies. Particularly, photodynamic therapy (PDT) consumes oxygen in treated tissues, which in turn lowers its efficacy. Tools for online monitoring of intracellular pO₂ are desirable.

Methods: The pO₂ changes were tracked during photodynamic therapy (PDT) with δ-aminolevulinic acid (ALA) in mouse skin, xenograft tumors, and human skin. ALA was applied either topically as Ameluz cream or systemically by injection. Mitochondrial pO₂ was quantified by time-gated lifetime-based imaging of delayed fluorescence (DF) of protoporphyrin IX (PpIX).

Results: pO₂-weighted images were obtained with capture-times of several seconds, radiant exposures near 10 mJ/cm², spatial resolution of 0.3 mm, and a broad dynamic range 1-50 mmHg, corresponding to DF lifetimes ≈20-2000 μs. The dose-rate effect on oxygen consumption was investigated in mouse skin. A fluence rate of 1.2 mW/cm² did not cause any appreciable oxygen depletion, whereas 6 mW/cm² and 12 mW/cm² caused severe oxygen depletion after radiant exposures of only 0.4-0.8 J/cm² and <0.2 J/cm², respectively. Reoxygenation after PDT was studied too. With a 5 J/cm² radiant exposure, the recovery times were 10-60 min, whereas with 2 J/cm² they were only 1-6 min. pO₂ distribution was spatially non-uniform at (sub)-millimeter scale, which underlines the necessity of tracking pO₂ changes by imaging rather than point-detection.

Conclusions: Time-gated imaging of PpIX DF seems to be a unique tool for direct online monitoring of pO₂ changes during PDT with a promising potential for research purposes as well as for comparatively easy clinical translation to improve efficacy in PDT treatment.

Keywords: ALA photodynamic therapy; Imaging of intracellular oxygen; Oxygen depletion and recovery; Protoporphyrin IX; Time-gated delayed fluorescence; pO(2) imaging.

Figure 1:

a) Jablonski diagram showing the origin of delayed fluorescence of PpIX from reverse intersystem crossing (RISC), b) experimental setup, c) Schematic representation of prompt fluorescence kinetics versus delayed fluorescence kinetics, d) comparison of DF kinetics in mouse skin in a live animal (pO₂ ~50 mmHg) and dead animal (pO₂ ~ 0 mmHg).

Figure 2:

(a) mouse #1 four hours after topical application of Ameluz: white-light image, PF image, and a combined image; (b) mouse #2, the cyan rectangle shows the area depicted in panels c and d; (c) mouse #1, column 1&2: first and fifth frame of the DF image sequence (the starting and ending gate-times are stated with respect to the end of the excitation pulse), col.3: DF lifetime image with a single-exponential fitting, col.4: pO₂* image with a single-exponential fitting, col.5: DF lifetime image with a two-exponential fitting, col.6: pO₂* image with a two-exponential fitting; (d) mouse #2; (e) comparison of DF kinetics from two individual pixels, with a single- and two-exponential fit for illustration (dashed and solid line, resp.); (f) overall DF kinetics summed over the whole irradiated area; (g) DF kinetics and pO₂* images (two-exponential fitting) at three different breathing rates.

Figure 3:

Mouse skin treated with Ameluz. (a) Integral DF intensity, DF lifetime, and pO₂* after increasing radiant exposures during PDT (0, 580, 670, and 720 mJ/cm²) in a spot irradiated at fluence rate of 6 mW/cm². The exp3 fit model was used. (b) The lifetime of DF kinetics (summed over the whole irradiated spot) gets longer with increasing radiant exposures during PDT. (c) Evolution of the average pO₂* during PDT with different fluence rates (6 mW/cm² blue, 1.2 mW/cm² cyan, 12 mW/cm² purple). The drop of pO₂ at 12 mW/cm² was so rapid that an accurate determination of the curve was not possible and therefore only the purple arrow is shown to indicate the steep decrease. The red curve shows that the integral DF intensity steeply grows as the pO₂* decreases at fluence rate of 6 mW/cm².

Figure 4:

(a) Oxygen recovery after PDT with a radiant exposure of 5 J/cm². The pO₂* increases during the time after the end of irradiation, but in some spots very low pO₂* remains even after > 20 minutes. In contrary, integral DF intensity is the largest right after PDT and gradually drops. (b) Another example of pO₂ recovery after radiant exposure of 5 J/cm². (c) Oxygen recovery after 2 J/cm² radiant exposure. The exp3 fit model was used for all pO₂* images.

Figure 5:

Imaging of pre-existing hypoxia in a mouse with a tumor 6hrs after intraperitoneal injection of 250 mg/kg ALA. (a) tumor covered with skin, (b) tumor exposed by cutting a window in the skin. The DF intensity is plotted in a log scale in b) in order to accommodate the broad dynamic range. The exp2 fit model was used.

Figure 6:

A human finger 4 hrs after topical application of Ameluz. The pO₂ was modulated by strangling the finger with rubber band, and subsequent release. The DF time-gated images are shown in a log scale to accommodate for the large dynamic range. The exp2 fit model was used to calculate the pO₂* images. The graph on the right shows comparison of DF kinetics from the central and peripheral area of the investigated spot.

Figure 7:

PF-lifetime-based imaging of PpIX photobleaching in mouse skin after application of Ameluz. (a) PF lifetime image of three spots before and after 50 J/cm² PDT (100 mW/cm²) was delivered to the central spot only. (b) Normalized PF kinetics from another mouse during PDT irradiation in the range 0–50 J/cm² with a 10 J/cm² increment. (c) Development of PF lifetime, normalized PF intensity, and relative contribution of the fast PF component as a function of radiant exposure during PDT. (d) Normalized PF spectra as a function of radiant exposure during PDT.