Sensitive simultaneous measurements of oxygenation and extracellular pH by EPR using a stable monophosphonated trityl radical and lithium phthalocyanine

Abstract

The monitoring of acidosis and hypoxia is crucial because both factors promote cancer progression and impact the efficacy of anti-cancer treatments. A phosphonated tetrathiatriarylmethyl (pTAM) has been previously described to monitor both parameters simultaneously, but the sensitivity to tackle subtle changes in oxygenation was limited. Here, we describe an innovative approach combining the pTAM radical and lithium phthalocyanine (LiPc) crystals to provide sensitive simultaneous measurements of extracellular pH (pH_e) and pO₂. Both parameters can be measured simultaneously as both EPR spectra do not overlap, with a gain in sensitivity to pO₂ variations by a factor of 10. This procedure was applied to characterize the impact of carbogen breathing in a breast cancer 4T1 model as a proof-of-concept. No significant change in pH_e and pO₂ was observed using pTAM alone, while LiPc detected a significant increase in tumor oxygenation. Interestingly, we observed that pTAM systematically overestimated the pO₂ compared to LiPc. In addition, we analyzed the impact of an inhibitor (UK-5099) of the mitochondrial pyruvate carrier (MPC) on the tumor microenvironment. In vitro, the exposure of 4T1 cells to UK-5099 for 24 hours induced a decrease in pH_e and oxygen consumption rate (OCR). In vivo, a significant decrease in tumor pH_e was observed in UK-5099-treated mice, while there was no change for mice treated with the vehicle. Despite the change observed in OCR, no significant change in tumor oxygenation was observed after the UK-5099 treatment. This approach is promising for assessing in vivo the effect of treatments targeting tumor metabolism.

Introduction

The reprogramming of energy metabolism of cancer cells is considered a hallmark of cancer [1–3]. The extracellular acidosis associated with the altered energy metabolism may have a profound impact on cancer progression, on immunosurveillance and on the efficacy of treatments [4–6]. In addition, hypoxia is a common feature of solid tumors that contributes to angiogenesis, invasiveness, metastasis, genomic instability and resistance to anticancer therapy (radiotherapy, chemotherapy and immunotherapy) [7, 8]. Importantly, it has been demonstrated that both extracellular acidosis and oxygen availability are not necessarily associated in solid tumors [9]. Having the capability to assess both extracellular pH (pH_e) and partial pressure of oxygen (pO₂) in tumors should lead to improvements in patient stratification regarding the selection of treatment and optimization of combined therapies [8, 10]. In this context, low-frequency Electron Paramagnetic Resonance (EPR) spectroscopy has gained interest for its capability to measure both tissue acidosis and oxygenation.

The measurement of pH in tissues using EPR can be achieved by measuring variation in hyperfine splitting constants in the EPR spectra of pH-sensitive imidazolidine nitroxides [11–13] or phosphonated trityl radicals [14, 15]. EPR has also been investigated for its capability to provide highly sensitive measurements of tissue oxygenation [16–20] thanks to the oxygen-induced changes in electron relaxation properties of paramagnetic oxygen sensors such as soluble stable free radicals (nitroxide or trityl radicals) [21, 22] or particulate materials (charcoals, carbon blacks, lithium phthalocyanine (LiPc) and derivatives) [23–27]. Several dual probes allowing the simultaneous characterization of parameters of the tumor microenvironment have been described [28–31]. For example, Ilangovan et al used LiPc crystals together with a nitroxide to measure tumor oxygenation and redox status during hyperoxygenation treatment [32]. An interesting development has been the synthesis of the monophosphonated tetrathiatriarylmethyl (pTAM or HOPE) radical that is able to provide simultaneous measurement of pH_e and pO₂ [33–35]. This asymmetric monophosphonated trityl radical displays a doublet pattern. The pH can be derived from the respective heights of the EPR lines (that depend on the molar fraction of the acidic form Pa versus basic form Pb) while the pO₂ can be measured from the linewidths (Figure 1a) [34].

Figure 1: EPR sensors used in the present study.

a) Structure of monophosphonated triarylmethyl radical (pTAM/HOPE) and its typical EPR spectrum showing both ionic forms (acidic form, Pa and basic form, Pb). The respective height of both forms provides information on the pH of the environment while the linewidth provides information on the oxygen concentration. Adapted from [33]

b) Structure of lithium phthalocyanine (LiPc) and its EPR spectrum. The variation in linewidth can be calibrated as a function of pO₂.

c) Typical EPR spectrum of both pTAM and LiPc recorded in vivo using a 1 GHz spectrometer. Note that the EPR spectra of pTAM and LiPc do not overlap.

While the multifunctionality of the probe represents a remarkable achievement, soluble paramagnetic sensors present an intrinsic limited sensitivity to tackle subtle variations in oxygenation that may occur in vivo. Because it is known that particulate materials display a much higher sensitivity to variations in partial pressure of oxygen [16], this prompted us to develop a new approach with a combination of pTAM together with lithium phthalocyanine crystals. We selected LiPc because this material presents an extremely narrow EPR signal and a high sensitivity to oxygen variations (Figure 1b) [25]. In addition, the EPR signal of LiPc does not interfere with the doublet of pTAM (Figure 1c). To assess the value of our approach, we first compared in vitro the sensitivity of pTAM and LiPc to variations in the oxygen environment. We then evaluated the value of pTAM alone and pTAM/LiPc in vivo in tumors using treatments that could potentially have an impact on the pO₂ and/or the pH_e. In a first step, we exposed mice to carbogen (95% O₂, 5% CO₂) breathing. This treatment is indeed often used in experimental studies as a positive control for its capability to increase tumor oxygenation [36–40]. In addition, acidification could potentially occur due to the CO₂ component of the inspired gas [41]. In a second step, we evaluated the effect of the inhibition of the mitochondrial pyruvate carrier (MPC). MPC is a heterodimeric complex located in the mitochondrial inner membrane that is responsible for importing pyruvate, the end-product of glycolysis, from the cytosol into the mitochondrial matrix (Figure 2). It was previously found that 7ACC2, an inhibitor of MPC, increased the extracellular acidification rate in squamous cervix carcinoma cells (SiHa) and alleviated tumor hypoxia through a decrease in oxygen consumption rate (OCR) in squamous hypopharyngeal carcinoma cells (FaDu) [42]. We also recently demonstrated, using CEST-MRI with iopamidol, that another inhibitor of MPC, UK-5099, induced acidification in a breast cancer model (4T1) [43]. Therefore, this model and this treatment were selected to evaluate the performances of the dual EPR probes. This was achieved in vitro by monitoring the effect of cell exposure to UK-5099 on extracellular acidification and oxygen consumption rate. We also monitored in vivo the effect of UK-5099 on pH_e and pO₂ in 4T1 tumor-bearing mice.

Figure 2: Role of the mitochondrial pyruvate carrier (MPC).

MPC is responsible for importing pyruvate, the end-product of glycolysis, from the cytosol into the mitochondrial matrix. Its inhibition may potentially lead to extracellular acidification and a decrease in oxygen consumption.

Materials & Methods

Reagents

The mono-phosphonated tetrathiatriarylmethyl radical was synthesized as previously described [34]. ¹⁵N-PDT (4-oxo-2,2,6,6-tetramethylpiperidine-d₁₆-¹⁵N-1-oxyl) was purchased from CDN Isotopes (Pointe-Claire, Canada). Lithium phthalocyanine crystals were provided by H.M. Swartz (Dartmouth Medical School, Hanover, NH, USA). UK-5099, dextran from Leuconostoc Mesenteroides (average MW 60000–76000) and dimethyl sulfoxide (DMSO) were from Sigma-Aldrich (Overijse, Belgium).

Cell Lines and Culture

4T1 breast cancer cells were cultured at 37°C in a humidified atmosphere with 5% CO₂ and maintained in RPMI medium (Thermo Fisher Scientific, Seneffe, Belgium) supplemented with 10% heat-inactivated FBS. The cell line was purchased from ATCC where it was regularly authenticated by short tandem repeat profiling. The 4T1 cell line was tested for mycoplasma contamination with the PCR-based MycoplasmaCheck assay (Eurofins, Ebersberg, Germany) before being used.

EPR instrumentations

In vitro studies were performed using a Bruker EMX-Plus spectrometer operating in X-band (9.85 GHz) and equipped with a PremiumX ultra-low noise microwave bridge and a SHQ high sensitivity resonator. The EPR cavity was heated to 310 K with continuous air flow during all experiments.

In vivo EPR spectra were recorded using an EPR spectrometer (electromagnet from Magnettech, Berlin, Germany; electronic console from Clin-EPR, Lyme, NH, USA) with a low-frequency microwave bridge operating at 1.1 GHz and an extended loop (surface coil) resonator (Clin-EPR). The sensitive volume under the loop is about 1 cm³.

pH_e and pO₂ calibrations

Saline solutions (containing 0, 1, 2, 3, 4 % albumin) containing either pTAM (200 μM) or LiPc were placed in a gas-permeable polytetrafluoroethylene (PTFE) tubing (inner diameter 0.025 in.; outer diameter 0.029 in.; Zeus Industrial Products, Letterkenny, Ireland). The oxygen content was varied between 0 and 21% O₂ using an Aalborg gas mixer. The oxygen content in the mixed gas was measured using a Servomex MiniMP 5200 oxygen analyzer (precision 0.1% oxygen content). pH was adjusted using NaOH or HCl. pH was measured using a pH meter. The experimental acquisition parameters for the pTAM and LiPc calibrations were as follows: microwave power, 2.518 mW; modulation frequency, 10 kHz; modulation amplitude, 0.0012mT; sweep width, 5 mT.

Oxygen Consumption Rate (OCR) and pH_e measurements in vitro

4T1 tumor cells were treated for 24 hours with either a vehicle (DMSO) or UK-5099 (10 μM). On the day of the experiment, cells were harvested to obtain 5 × 10⁶ cells/mL. The experimental mixture was prepared with 60 μL of cells, 40 μL of dextran (20%) and pTAM (200 μM. Parallel experiments were also done using ¹⁵N-PDT (2 mM) as oxygen sensor instead of pTAM because this nitroxide has been classically used in OCR experiments measured by EPR [44, 45]. This mixture was placed in a gas-impermeable hematocrit capillary sealed with gum, inserted into a quartz tube and placed in the heated cavity (310 K). The experimental parameters were the following: microwave power, 2.518 mW; modulation frequency, 10 kHz; modulation amplitude, 0.0012mT; sweep width, 5 mT. EPR spectra were recorded every other minute during 20 minutes for linewidth determination. As the linewidth reports on pO₂, oxygen consumption rates were obtained by measuring the pO₂ in the closed tube over time and determining the slope of the resulting linear plot (Figure S1). Full detailed practical procedure has been published elsewhere [46].

Tumor models and animal procedures

All experiments involving animals were performed in agreement with the Belgian law concerning the protection and welfare of the animals and were approved by the UCLouvain ethics committee (Agreement reference: 2022/UCL/MD/047). 2 × 10⁵ 4T1 cells in 100 μL of PBS were injected intramuscularly in the right hind paw of 6-week-old female BALB/cAnNRj mice.

LiPc crystals were implanted inside the tumor tissue of anesthetized animals 24 hours before the first measurement when tumors reached a volume of 200 ± 50 mm³. A small incision was made in the hind paw to insert 3–4 LiPc crystals. The small wound was sutured after insertion of the sensors. 50 μL of pTAM (3 mM in saline) was injected inside the tumor 3 minutes before positioning the mouse in the center of the 1 GHz spectrometer and launching the first EPR acquisition. Acquisitions parameters were as follows: power, 7.94 mW; modulation frequency, 3.29 kHz; modulation amplitude, 0.04 G; scan range, 3 G; scan time, 30 s; 10 acquisitions).

Animals were anesthetized by inhalation of isoflurane mixed with air (21% oxygen) in a continuous flow (2 l/min) delivered by a nose cone. The induction of anesthesia was performed with 3% isoflurane which was then stabilized at 1.5% for a minimum of 15 min before any measurement. It was previously demonstrated that this anesthesia regimen did not disturb the hemodynamics in mice [47]. A circulating water system was used for body-temperature regulation at 37 °C.

In vivo manipulation of pO₂ and/or pH_e

Carbogen Breathing:

50 μL of pTAM (3 mM in saline) was injected inside the tumor 3 minutes before the first EPR acquisition when mice were breathing air. After the first acquisitions, mice were kept in place but the gas was switched to carbogen (95% O₂, 5% CO₂). EPR spectra were acquired after 10 minutes of carbogen breathing.

MPC inhibition:

The treatment administration scheme was based on a previous established to monitor the pH_e changes induced by UK-5099 changes using CEST-MRI in the same tumor model [43]. Mice were treated with daily intraperitoneal injection of UK-5099 (3 mg/kg) or vehicle for 4 days. The basal pH_e and the basal pO₂ were measured on day first before the administration of the first dose of UK-5099. The measurement post-treatment was performed on day 4, one hour after the last injection of UK-5099. Control experiments were done using the vehicle (DMSO) instead of UK-5099.

The in vivo protocols are summarized in Figure 3.

Figure 3: Experimental protocols applied in vivo with simultaneous assessments of pO₂ and pH_e.

a) Protocol of carbogen breathing

b) Protocol of MPC inhibition using UK-5099

Results

LiPc is 10 times more sensitive to oxygen changes than pTAM

The in vitro calibration curve (Pa fraction as a function of pH) using pTAM is shown in Figure 4a. We also performed a direct comparison of the changes in linewidth measured for LiPc and pTAM (at different concentrations of albumin) as a function of pO₂ variation (Figure 4b). The comparison of the slopes showed that the sensitivity of LiPc was 10 times higher than that of pTAM to variations in the oxygen environment. The presence of albumin did not significantly change the response to variations in oxygenation as all calibration curves obtained were superimposable.

Figure 4: Calibrations of the EPR sensors.

a) pH_e calibration (ratio Pa/Pb as a function of pH) of pTAM

b) pO₂ calibrations (linewidth as a function of pO₂) of LiPc (in saline) and pTAM (in saline and 1-2-3-4 % albumin).

LiPc, not pTAM, detects significant changes in pO2 after carbogen breathing

As the EPR signal of pTAM decreased rapidly in tumors (T_1/2 = 13 min), measurements were done rapidly after pTAM administration. The carbogen challenge was used in order to increase the oxygenation and possibly modify the pH in the tumors. We did not observe any significant change in pH_e after carbogen breathing (Figure 5a). The results obtained for the evolution of tumor oxygenation were different between the pTAM and the LiPc oxygen sensors. While pTAM did not detect a significant change in tumor pO₂ after carbogen breathing (Figure 5b), LiPc indicated that breathing carbogen led to a significant increase in pO₂ in the same tumors (Figure 5c). Intriguingly, as shown in Figure 5d, the pO₂ readings obtained with pTAM were systematically and significantly higher than those measured with LiPc.

Figure 5: Effect of carbogen breathing on the tumor pH_e and pO₂ measured simultaneously.

a) pH_e measured before and after 10 min of carbogen breathing (n=4)

b) pO₂ measured before and after 10 min of carbogen breathing using the calibration of pTAM (n=4)

c) pO₂ measured before and after 10 min of carbogen breathing using the calibration of LiPc (n=4)

d) Individual comparison between pO₂ readings obtained from LiPc and pTAM from the same tumors

*, p < 0.05; **, p < 0.01, Student’s t-test. Note that pH_e does not seem altered by the treatment with carbogen. Note that pTAM was not able to tackle any significant variation in pO₂ while LiPc suggested a significant increase in pO₂. The pO₂ readings obtained with pTAM were systematically and significantly higher than those measured with LiPc.

pTAM detected changes in extracellular pH and in OCR in vitro in cells exposed to UK-5099

Oxygen consumption was monitored in 4T1 cells exposed for 24 hours to either DMSO (vehicle) or 10 μM of UK-5099. This was achieved using pTAM (Figure 6a) or ¹⁵N-PDT (Figure 6b), the latter probe being classically used for OCR measurement using EPR [44–46]. Both probes similarly revealed that the exposure to UK-5099 induced a decrease in OCR. The extracellular pH was also monitored in vitro using pTAM revealing that the pH_e was significantly decreased after UK-5099 exposure (Figure 6c).

Figure 6: Effect of UK-5099 exposure for 24 hours on oxygen consumption rate and extracellular pH.

a) OCR measured using pTAM

b) OCR measured using ¹⁵N-PDT

c) pH_e measured using pTAM

*, p < 0.05, Student’s t-test.

UK-5099 treatment induced changes in pH_e but not in pO₂ in vivo in 4T1 tumor models

Mice bearing 4T1 tumors implanted in the leg were treated for 4 days either with DMSO (vehicle) or UK-5099 (3 mg/kg). This administration pattern was inspired by our previous work with UK-5099 which showed a significant effect on pH_e values when measured by CEST-MRI [43]. The results obtained in vivo showed a significant decrease in tumor pH_e (by 0.39 pH units) in UK-5099-treated mice while there was no significant change over time for mice treated with the vehicle (Figure 7a,b). The oxygenation was also monitored using the EPR linewidths of pTAM or LiPc. Both readings indicate no significant change in tumor oxygenation after vehicle administration or UK-5099 treatment for 4 days (Figure 7c–f). Again, we observed that pO₂ readings obtained with pTAM were much higher than those obtained using LiPc. For example, the pO₂ estimated before UK-5099 treatment was 89.20 ± 13.14 mmHg (mean ± SE) using pTAM while it was 6.231 ± 1.757 mmHg (mean ± SE) using LiPc.

Figure 7: Effect of UK-5099 treatment on pH_e and pO₂ measured in vivo in 4T1 tumor models.

a) pH_e measured before and after 4 days of DMSO treatment (n=7)

b) pH_e measured before and after 4 days of UK-5099 treatment (3 mg/kg) (n=10)

c) pO₂ measured before and after 4 days of DMSO treatment as estimated by pTAM (n=7)

d) pO₂ measured before and after 4 days of UK-5099 treatment (3 mg/kg) as estimated by pTAM (n=11)

e) pO₂ measured before and after 4 days of DMSO treatment as estimated by LiPc (n=6)

f) pO₂ measured before and after 4 days of UK-5099 treatment (3 mg/kg) as estimated by LiPc (n=8)

****, p < 0.0001, Student’s t-test.

Discussion

Acidosis and hypoxia are features often observed in malignant tumors. Both factors are considered as poor prognostic markers because their presence affects the curability of solid tumors [48, 49]. In this context, assessing both pO₂ and pH_e is of paramount importance not only to evaluate the aggressiveness of a tumor but also to select treatment options. Indeed, intense research is currently under development to define therapeutic interventions that may modulate the tumor metabolism and tumor microenvironment to optimize anti-cancer strategies [4, 50–54]. Innovative EPR sensors have been developed with the capability to monitor simultaneously tissue oxygenation and acidity [28, 31, 33–35, 55–58] with interesting applications for tumor profiling [59]. Unfortunately, the sensitivity of soluble probes to variations in tissue oxygenation is limited. Our results obtained in vivo after carbogen breathing illustrate this limited sensitivity of pTAM (Figure 5). Oxygen-enriched gas breathing (including pure oxygen, carbogen or hyperbaric oxygen) has been used in many experimental studies (including EPR oximetry studies) to manipulate tumor oxygenation and alleviate tumor hypoxia [36–40, 60]. In the present study, the soluble probe pTAM did not detect significant changes in pO₂ in sharp contrast to readings done with the particulate probe LiPc. To understand the limited sensitivity of soluble probes, it is worth looking back at the principles of EPR oximetry [16, 18, 61]. Bimolecular collisions between paramagnetic oxygen and free radicals are at the origin of changes in EPR spectra because they modify the resonance characteristics of the radical and consequently the EPR spectrum. The enhancement of relaxation rates scales linearly with the concentration of oxygen. For soluble probes, there is a direct relationship between the EPR line exchange broadening and the radical-radical collision rate [60] The exchange rate ω is governed by the Smoluchowski equation:

$$\omega =4\pi R\left\{\text{D}\left({\text{O}}_2\right)+\text{D}\left(\text{SP}\right)\right\}\left[{\text{O}}_2\right]$$

and the variation of the EPR linewidth Δ LW is

$$\Delta \text{LW}=4\pi Rkp\left\{\text{D}\left({\text{O}}_2\right)+\text{D}\left(\text{SP}\right)\right\}\left(\Delta \left[{\text{O}}_2\right]\right)$$

where R is the interaction distance, k is a proportionality constant, p is the probability that exchange will occur upon each collision, D(O2) and D(SP) are the diffusion coefficients of oxygen and the spin probe, respectively, and [O₂] is the concentration of oxygen in solution. In practice, it means that, for small nitroxides spin probe in water solution, the variation in Δ LW when shifting from nitrogen (0 % O₂) to air (21 % O₂) will be around 10 μT This value actually depends on the R parameter as bulky radicals present a smaller oxygen sensitivity (8.5 μT for dFT (deuterated Finland trityl), 5.6 μT for OXO63 (Finland trityl substituted with -CH₂CH₂OH), 3.0 μT for SOX71 (succinylated derivative of Finland trityl OXO71), as illustrative examples) [62]. Of note, changes in oxygen concentrations in tissues are generally ranging from 0 to 7 % O₂. In contrast, the sensitivity in linewidth variation for particulate material is much larger, with a broadening per unit of pO₂, exceeding that of soluble probes by several orders of magnitude [16]. While the mechanism of the relationship between Δ LW and Δ pO₂ is not fully elucidated for all particulate probes [63], it has been described that, for LiPc, the oxygen sensitivity is critically dependent on the crystal forms. Oxygen has been shown to migrate into the channels presented by the tetragonal x-form where it influences the magnetic properties of the delocalized free electrons present in LiPc [64, 65]. The sensitivity measured for LiPc crystals used in the present study was comparable to that described in its first application in vivo [25]. As described in Figure 4b, the sensitivity to oxygen variations was ten times higher for LiPc than for pTAM. This material is particularly interesting because its EPR spectrum does not overlap with that of pTAM. While other particulate materials such as charcoals and carbon blacks present a larger sensitivity to variations in oxygenation, their EPR spectra are too large in oxygenated media for being applied together with pTAM due to the overlapping of the EPR lines.

As pTAM and LiPc were simultaneously present in tissues, it offered the possibility to compare the pO₂ values obtained by both sensors. In both experiments (carbogen challenge or MPC inhibition), we systematically observed that apparent pO₂ values obtained from the linewidth of pTAM were significantly larger than those obtained with LiPc (Figures 5d and 7d–f). Large pO₂ values using pTAM were also reported recently in a spontaneous breast cancer model [59]. Of note, most readings obtained with LiPc provided pO₂ values lower than 20 mmHg, data consistent with several dozens of studies carried out in tumors using EPR oximetry and Clark electrodes, as reviewed in Gallez and Vaupel et al. [66, 67]. While remaining speculative, several factors may contribute to this discrepancy in pO₂ estimates by both probes. Contrarily to LiPc, the EPR linewidth of pTAM is affected by “concentration effects”. Indeed, a high local concentration of the probe induces a self-broadening of the EPR linewidth that will lead to an overestimation of the actual pO₂. The concentration used for the injection was 3 mM which was the highest concentration where no self-broadening of the probe was observed in vitro (Figure S2). An increase in the linewidth and a loss in hyperfine resolution was observed at higher concentration (5 and 10 mM) (Figure S2). Of note, others have used a concentration of 2 mM for injection of pTAM [68]. As the probe is diluted in the biological media after injection, in principle, we could not expect a self-broadening of the EPR signal. However, we cannot exclude that the probe could concentrate after injection in a specific location leading to a self-broadening. To avoid this concentration effect observed with soluble probes, it has been suggested that measuring the T1 relaxation times should be preferred to T2, a possibility offered by pulsed EPR spectrometers [69]. Another potential factor that could lead to a broadening of the signal is the binding to macromolecules. It has been described that pTAM can bind to albumin [35]. The binding to albumin leads to a significant decrease in the height of the EPR signal (Figure S3). It is likely that the EPR spectrum is the combination of a narrow signal (due to free pTAM) and a very broad signal due to bound pTAM to albumin. When analyzing the linewidth of pTAM in the absence or in the presence of albumin (up to 4%, the physiological plasma concentration), we did not observe a larger EPR linewidth for pTAM in the presence of albumin (Figure S3b). Most importantly, the calibration curves obtained using different concentrations of albumin (from 0 to 4%) did not reveal any change in sensitivity to variations in oxygenation (Figure 4b). It is likely that the very low modulation amplitude and modulation frequency used for pO₂ and pHe measurements in our study makes the contribution of the broad signal negligible in the EPR spectra. Finally, it should be noted that both probes could interrogate different areas of the experimental tumors. LiPc crystals were implanted one day before the EPR spectroscopy experiment. The volume sampled by the LiPc crystals is limited (around 1–2 mm³). The solution of pTAM was injected just before the acquisition of the EPR spectrum. Being soluble, pTAM may diffuse more largely in the tumor and we can hypothesize that the probe could access to areas better perfused and therefore with higher pO₂ [59].

Using carbogen breathing, we demonstrated the highest value of the association with LiPc to tackle changes in tissue oxygenation. Regarding the pH_e estimation, carbogen breathing did not lead to significant variation in acidity (Figure 5a). It is likely that the period of gas breathing was too short to lead to a decrease in pH_e due to the dissolution of CO₂ in the blood. Therefore, to illustrate the dual capability of measuring both pO₂ and pH_e simultaneously, we used a treatment that targets MPC. By blocking the import of pyruvate into the mitochondria, we could expect an increase in extracellular acidity (by blocking lactate uptake and favoring lactate export outside the cell) and a decrease in oxygen consumption rate (by blocking the pyruvate-fueled respiration) [42]. In vitro, pTAM detected both extracellular acidification (Figure 6c) and a decrease in OCR (Figure 6a). Of note, we previously reported a change in extracellular acidification rate in the same model using the Seahorse XF technology [43], and the results obtained with pTAM regarding the change in OCR were consistent with those observed with the ¹⁵N-PDT probe used as a standard for this type of measurement [44–46]. Knowing that the model was responsive to UK-5099 treatment in vitro, it was rational to translate these experiments in vivo. In animals receiving UK-5099 injection (3mg/kg) during four days, we observed extracellular acidification, a result that is consistent with our previous study using CEST-MRI and iopamidol as contrast agent [43]. Contrary to our expectations, we did not observe any significant change in tumor oxygenation (Figure 7f). This observation seems counterintuitive as many treatments that decrease tumor cell metabolism and oxygen consumption have been shown to alleviate tumor hypoxia (as reviewed in [50]). Our hypothesis is that perfusion effects could counteract the effect on oxygen consumption. As the treatment was applied for four days, the tumors continued to grow. As the pO₂ was measured after four days of treatment, we could assume that a decrease in perfusion could have occurred, masking the effect on oxygen metabolism. At this stage, this remains a hypothesis that would require further investigation.

Conclusions

We have shown that the combination of a phosphonated tetrathiatriarylmethyl (pTAM) radical with LiPc crystals allows simultaneous quantitative measurement of pO₂ and pH_e. While the use of the single probe pTAM alone allows the measurement of pH_e and pO₂ from the same site in biological media, this probe suffers from limited sensitivity to oxygen variations. The main advantage of this combination over the use of pTAM alone is the gain in sensitivity to oxygen variations (by a factor of 10). This dual-sensor has been successfully applied in animal models. Incidentally, we also observed that pTAM systematically provided higher pO₂ values than those obtained with LiPc. This new method seems promising for studying the effect of treatments targeting tumor metabolism in vivo.