Tissue acidosis does not mediate the hypoxia selectivity of [64Cu][Cu(ATSM)] in the isolated perfused rat heart

Copper-64-Diacetyl-bis(N4-methylthiosemicarbazone) [64Cu][Cu(ATSM)] is a hypoxia-targeting PET tracer with applications in oncology and cardiology. Upon entering a hypoxic cell, [64Cu][Cu(II)(ATSM)] is reduced to a putative [64Cu][Cu(I)(ATSM)]− species which dissociates to deposit radiocopper, thereby providing hypoxic contrast. This process may be dependent upon protonation arising from intracellular acidosis. Since acidosis is a hallmark of ischemic tissue and tumors, the hypoxia specificity of [64Cu][Cu(ATSM)] may be confounded by changes in intracellular pH. We have therefore determined the influence of intracellular pH on [64Cu][Cu(ATSM)] pharmacokinetics. Using isolated perfused rat hearts, acidosis was induced using an ammonium pre-pulse method, with and without hypoxic buffer perfusion. Cardiac [64Cu][Cu(ATSM)] pharmacokinetics were determined using NaI detectors, with intracellular pH and cardiac energetics monitored in parallel by 31P NMR. To distinguish direct acidotic effects on tracer pharmacokinetics from acidosis-induced hypocontractility, parallel studies used lidocaine perfusion to abolish cardiac contraction. Hypoxic myocardium trapped [64Cu][Cu(ATSM)] despite no evidence of it being acidotic when characterised by 31P NMR. Independent induction of tissue acidosis had no direct effect on [64Cu][Cu(ATSM)] pharmacokinetics in either normoxic or hypoxic hearts, beyond decreasing cardiac oxygen consumption to alleviate hypoxia and decrease tracer retention, leading us to conclude that tissue acidosis does not mediate the hypoxia selectivity of [64Cu][Cu(ATSM)].

acidosis may alter the redox potential of the oxidised Copper (II) species, thereby altering the lipophilicity of the reduced Copper (I) species, or the dissociation kinetics of the complex 19 . In vitro UV-Vis spectroscopy studies have demonstrated that the stability of various [ 64 Cu][Cu(BTSCs)], such as PTSM and KTS complexes and their reduction products decrease substantially with reduced pH 11,19 , making them more prone to dissociation, while cyclic voltammetry suggests that [ 64 Cu][Cu(ATSM)] is more readily reduced in acidic conditions 19 . Ischemic myocardium and the tumour microenvironment are both potentially associated with acidosis. In myocardium, the net hydrolysis of ATP during ischemia leads to an increase in intracellular H + concentration 20 , which has been demonstrated by 31 P NMR spectroscopy to be more severe than that caused by hypoxia alone 21 . Commonly associated with hypoxia, acidosis is also a hallmark of aggressive tumors, driven by a glycolytic phenotype, the increased production and extrusion of H + and lactate, and limited perfusion 22 . The importance of acidosis is of increasing interest as a driver of tumor progression and metastatic spread 22 . If the hypoxia sensitivity of the [ 64 Cu][Cu(BTSC)] complexes is mediated by (or heavily modified by) changes in intracellular or intra-tissue pH, this would greatly complicate the use and interpretation of the imaging data that they provide. While in vitro studies and in silico modeling 19 suggest that pH may influence the hypoxia selectivity of these complexes, the issue has not yet been specifically investigated in a biologically relevant model of tissue hypoxia.
We have established an isolated perfused heart system coupled with a triple NaI gamma detection apparatus which allows the characterization of radiotracer pharmacokinetics in an intact functioning organ over which we have complete functional control 15,23 . Interventions can be performed accurately and reproducibly without the added complications of circulating tracer metabolites, under conditions which may otherwise be lethal in vivo. In this study, we have used this approach to investigate the interplay between tissue hypoxia and intracellular pH, independently and in combination, on radiotracer kinetics to establish whether tissue acidosis is fundamental to, or impacts upon, the hypoxia selectivity of [ 64 Cu][Cu(ATSM)].

Results
Cardiac hemodynamics. Cardiac  Cardiac lactate release. Lactate release from hearts in all treatment groups are summarized in Fig. 4.
Lactate release from normoxic control hearts was minimal during normoxic perfusion (0.085 ± 0.01 to 0.011 ± 0.03 mmol/L). Perfusion with hypoxic buffer caused elevated lactate washout, reaching a maximum of 0.35 ± 0.03 mmol/L after 20 mins of hypoxia. In normoxic hearts, perfusion with NH 4 Cl caused a transient but not significant small increase in lactate release from 0.10 ± 0.03 to 0.15 ± 0.11 mmol/L within 2 minutes which decreased again upon NH 4 Cl washout. Lactate release from acidotic hearts perfused with hypoxic buffer did not increase throughout the protocol and were not significantly different from non-acidotic control hearts. Lidocaine infusion abolished lactate release during normoxic buffer perfusion, which was not significantly different from the low lactate washout observed in the acidotic group, but significantly lower than that observed in the hypoxia alone perfusion group (p < 0.05).
Cardiac Copper-64 retention. Representative traces from the Na/I detector interrogating the heart during examples of each perfusion protocol are shown in Fig. 5, and summarized across all hearts in Fig. 6. During normoxic perfusion, 10.1% ± 2.5% of the injected dose (ID) was retained in the heart 20 min after the first injection, and this level of radiotracer retention was consistently observed in the subsequent injections in the normoxic control heart. Induction of acidosis in normoxic hearts caused a decrease in cardiac Copper-64 retention from 9.3% ± 2.7% during the control period to 2.4% ± 1.6% (p < 0.05), returning to 12.7% ± 3.0% 15 mins after zoniporide/acid washout. Perfusion with hypoxic buffer caused a significant increase in tracer retention from 11.0% ± 2.2% to 46.5% ± 12.0% ID (p < 0.05). However, inducing acidosis in hypoxic hearts abolished Copper-64 retention such that it was not significantly different to pre-hypoxic values. Lidocaine infusion in normoxic hearts had a similar effect to acidosis, with Copper-64 retention falling from 9.22% ± 0.79% to 4.93% ± 0.8% (p < 0.05) before returning to pre-lidocaine values (10.7% ± 0.6% ID). Lidocaine infusion also abolished the elevated Copper-64 retention previously seen in hypoxic hearts.   ] retention was lower than in untreated normoxic control hearts. While the Langendorff isolated perfused heart is energetically stable for several hours, it is potentially at the brink of normoxia because KHB has a lower oxygen-carrying capacity than blood, even when saturated with 95% O 2 /5% CO 2 . This is apparent from the significant vasodilation and high coronary flows associated with crystalloid versus blood perfused preparations 28 . Isolated hearts perfused with KHB are almost maximally vasodilated, which has led to the suggestion that they may be slightly hypoxic, particularly in the endocardium, where energy (and oxygen) demands are highest 29 . While in previous studies we had attributed the baseline level of [ 64 Cu][Cu(ATSM)] cardiac retention in "normoxically-perfused" isolated hearts to be due to non-specific retention in cell membranes (due to their lipophilicity), our data suggest that some of this retention may represent a small fraction of hypoxic cells within a crystalloid-perfused heart, which become normoxic once when the heart is mechanically unloaded. We have recently shown by 31 P NMR that mechanically uncoupling isolated hearts with blebbistatin elevates both phosphocreatine levels and ATP content in "normoxic" crystalloid buffer-perfused isolated hearts which would be consistent with this 30 . We exploited the ammonium prepulse approach to specifically induce intracellular acidosis, as opposed to perfusing with acidotic buffers, which would not necessarily translate to intracellular acidosis. This also allowed us to exclude the possibility of protonation of the complex by acidotic buffer before it reached the heart in our study. The ammonium prepulse approach coupled with NHE inhibition is an established means of inducing and maintaining intracellular acidosis in both basic cardiac research and other applications 31,32 ; similar, but less severe intracellular acidosis of pH 6.7 has previously been demonstrated with a similar protocol in isolated perfused ferret hearts using the NHE inhibitor 5-(N-ethyl-N-isopropyl)amiloride (EIPA) 33 , while similar effects on contractile function to those we report here have previously been demonstrated with NH 4 Cl infusion and NHE-1 inhibitor cariporide 34 . While NHE inhibition is associated with a loss of Ca 2+ uptake via NCX, acidosis-mediated loss of contractile function is not due to the decrease in Ca 2+ concentration, but to a decrease in contractile protein responsiveness to Ca 2+ 35 . We demonstrate that zoniporide treatment alone had no effect on Copper-64 retention during normoxia or hypoxia compared to vehicle control, which confirms that there was no interaction between zoniporide and [ 64 Cu][Cu(ATSM)] which may have affected myocardial uptake or dissociation.

Conclusion
Radiocopper bis(thiosemicarbazone) complexes represent a versatile family of hypoxia imaging agents with a range of hypoxia selectivities for a variety of applications in both cardiology and oncology. To optimize the diagnostic and prognostic insight gained from the PET images with these complexes, it is essential to understand the nature of their tissue uptake and retention. Here, we demonstrate that their hypoxia-dependent tissue retention is not dependent upon intracellular acidosis (nor indeed directly affected by it) and confirm their specificity to changes in intracellular oxygen saturation.

Materials and Methods
Reagents and gas mixtures. All reagents were purchased from Sigma Aldrich (Poole, Dorset, UK) unless otherwise stated. All gas mixtures were purchased from BOC, UK. Specialist gas mixtures were certified by the manufacturer.
[  intervention. An ammonium pre-pulse technique was employed to induce intracellular acidosis 38 , which was maintained by infusion of the sodium hydrogen exchanger (NHE-1) inhibitor zoniporide 39 in bicarbonate free buffer (to limit buffering capacity). Hypoxia was induced by perfusing hearts with KHB gassed with 20% O 2 .
Since tissue acidosis inhibits cardiac contractility, which would potentially have an oxygen-salving effect which may itself affect hypoxia tracer pharmacokinetics, we perfused further groups of hearts with 0.8 mM lidocaine to inhibit cardiac contractility to establish the effect of tissue hypocontractility on tracer pharmacokinetics independent of intracellular pH.

P MR Spectroscopic Analysis.
Changes in cardiac energetics and intracellular pH were monitored in real time in parallel groups of hearts using 31 P NMR spectroscopy. Hearts were cannulated and perfused with KHB in Langendorff mode and inserted into a 15 mm glass MR tube, which was inserted into a custom-built MR spectroscopy probe, as previously described 40 . 31 P NMR spectra were acquired on an Bruker Avance III 9.4 T spectrometer using a 15 mm 31 P/ 1 H birdcage coil 41 . Shimming was performed on the 1 H line shape of water (full width at half maximum <20 Hz). 31 P spectra were acquired with a pulse-acquire sequence using a 60° flip angle, a repetition time of 3.8 s, and 64 scans (4 min per spectrum). The peak area of each metabolite was normalized to that of phosphocreatine during normoxia.
Lactate measurement. 0.5 mL of coronary effluent was collected at regular intervals during perfusion and analyzed for lactate content to identify the onset of anaerobic glycolysis using a 2300 STAT Plus lactate analyzer (YSI Ltd) as described previously 37 .
Radiometric measurement. Cardiac radiotracer injection, retention and washout was monitored throughout using three orthogonally-arranged lead-collimated Na/I γ-radiation detectors (Raytest Isotopenmessgeräte GmbH, Straubenhardt, Germany) measuring Copper-64 activity at the input (arterial) perfusion line, the heart and the output perfusion line respectively. The detectors were connected to a Gina Star ™ data acquisition system (Raytest Isotopenmessgeräte GmbH) as previously described 15 . Data were normalized to the maximum peak counts after each injection and corrected for decay and cardiac background activity 30 seconds prior to each injection.
Statistical analysis. All data are presented as mean ± standard deviation. Statistical significance was calculated using a one-way ANOVA followed by Bonferroni post hoc test, or Dunnett test when multiple comparisons were made to a control group, using GraphPad Prism (GraphPad software Inc., San Diego, CA, USA).

Data Availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.