Assessment of tissue biochemical and optical scattering changes due to hypothermic organ preservation: a preliminary study in mouse organs

Ten balb/c female mice used in this study were 6–10 weeks old. The study was performed under the approval of the Research Ethics Committee of University College Cork. All methods were performed in accordance with the relevant guidelines and regulations. The mice were euthanized using cervical dislocation prior to the measurements. Subsequently, the mice were transported to the research center (Tyndall National Institute, Cork, Ireland) where access to their internal organs was enabled by (a) initial incision below the navel going up toward the mouth, (b) downward incision toward the tail, (c) lateral incision at the shoulder joints, (d) lateral incision at the pelvic girdle and (e) removal of the abdominal membrane. Once the internal organs could be accessed, they were rearranged in order to allow exposure of sufficiently large tissue surface for DRS data collection. Then, the mice were positioned at a metallic container and placed in a water bath kept at 37 °C until thermal equilibrium was reached. The temperature of all organs was monitored with a conventional infrared thermometer until their temperature reached between 30 °C and 37 °C (referred as body temperature or BT in this study). Once all DRS measurements were acquired at BT (three measurements per organ per mice), the mice were placed in a container with cold water with ice (0 °C) and left until thermal equilibrium was reached (0 °C–8 °C). Subsequently, three DRS measurements were taken per organ at cold temperature (CT). In total, we have evaluated ten hearts, ten kidneys, ten livers, and ten lungs at BT and CT. Figure 1 shows a schematic of the steps of organ collection and measurement protocol.

Zoom In Zoom Out Reset image size

The DRS system consisted of a tungsten halogen light source (350–2400 nm, HL-2000-HP, Ocean Optics, Edinburgh, United Kingdom), a one-to-four fan out low-OH silica 600 µm core fiber optic bundle (BF46LS01, Thorlabs, Munich, Germany) with 630 µm source-to-detector center-to-center distance, a visible/NIR spectrometer (350–1137 nm, QE-Pro, Ocean Optics, Edinburgh, UK) and a NIR spectrometer (1095–1919 nm, NIR-Quest, Ocean Optics, Edinburgh, UK). In this experimental setup, the broadband light is delivered to the tissue by using the illumination fiber of the quadrifurcated probe. The reflected light is collected by two fibers adjacent to the illumination fiber and captured by each spectrometer. The overlapping wavelength range (1095–1137 nm) was used to merge the two spectra into one, as explained in section 2.4. More details regarding the DRS system can be found elsewhere [32].

Prior to and after tissue measurements, background and reference data was collected for subsequent calculation of the tissue reflectance (section 2.4). The fiber optic probe was connected to a closed holder with black non-reflective walls and was kept at a fixed distance between the probe and a reflectance standard (FWS-99-01c, Avian Technologies LLC, New London, USA). The light source was turned on for reference measurements and turned off for background measurements. Tissue measurements were recorded by keeping the probe at 90° and in contact with the tissue surface. All measurements were collected using the same fiber optic probe. In total, 240 measurements were collected (30 per organ per temperature point).

This study used home-made MATLAB routines (MathWorks Inc., Natick, Massachusetts) for data preprocessing and analysis discussed in sections 2.4–2.6. Tissue reflectance $R\left( \lambda \right)$ was calculated as described by the equation (1):

$$\begin{equation}R\left( \lambda \right) = \frac{{{I^{{\text{Tissue}}}}\left( \lambda \right) - {I^{{\text{Background}}}}\left( \lambda \right)}}{{{I^{{\text{Reference}}}}\left( \lambda \right) - {I^{{\text{Background}}}}\left( \lambda \right)}},\end{equation} \tag{ 1 }$$

where ${ }{I^{{\text{Background}}}}$ is background intensity, ${I^{{\text{Tissue}}}}$ is intensity of signal collected from tissue, ${I^{{\text{Reference}}}}$ is reference intensity, and λ represents each wavelength within the wavelength range measured by a given spectrometer.

Once the reflectance was calculated for both visible/NIR (${R_{{\text{VIS}}/{\text{NIR}}}}\left( \lambda \right)$) and NIR spectrometers (${R_{{\text{NIR}}}}\left( \lambda \right)$), the spectrum at the overlapping wavelengths was used to calculate the area under the curve to normalize ${R_{{\text{NIR}}}}\left( \lambda \right)$. Next, the overlapping spectral region of ${R_{{\text{VIS/NIR}}}}\left( \lambda \right)$ and the normalized NIR spectrum ${R_{{\text{NIR}}\_{\text{norm}}}}\left( \lambda \right)$ were interpolated by using 100 points. Then, the spectra were merged in order to obtain the broadband DRS ${R_{{\text{exp}}}}\left( \lambda \right)$ according to the weighted sum as per equation (2):

$$\begin{equation}{R_{{\text{exp}}}}\left( \lambda \right) = \sum\limits_{i = 0}^{100} {\frac{{\left( {\left( {100 - i} \right) \times {\text{ }}{R_{{\text{VIS/NIR}}}} + i \times {\text{ }}{R_{{\text{NIR}}\_{\text{norm}}}}} \right)}}{{100}}} .\end{equation} \tag{ 2 }$$

After calculating the experimental reflectance ${R_{{\text{exp}}}}\left( \lambda \right)$, the DRS spectra were fitted in order to obtain tissue scattering properties (reduced scattering amplitude $\alpha {^{^{\prime}}}$, Mie scattering power b_Mie, and the percentage contribution f_Ray), and chromophore concentrations of blood (THb and THC), lipid (f_lipid), water (f_water), bile (f_bile), met-hemoglobin (f_metHb) as well as oxygen saturation (StO₂), and average vessel diameter R. These scattering properties and chromophore concentrations are the optimization variables fitted in the algorithm and determine the tissue absorption and scattering wavelength-dependent spectra. Combinations of absorption and scattering coefficients are associated to theoretical reflectance values obtained via a MC LUT obtained with MC simulations of light propagation in homogeneous media. This association allows the reconstruction of a theoretical DRS spectrum which is compared to the experimental spectrum until convergence of the spectral fitting and determination of the scattering properties and chromophore concentrations. Details of our spectral fitting algorithm can be found in the supplementary material (available online at stacks.iop.org/JPD/54/374003/mmedia). We built and used a MC LUT which covered µ_a values from 0.01 to 300 cm⁻¹ and µ_s from 0.1 to 1000 cm⁻¹ (spanning optical properties of most tissues [33, 34]) for our spectral fitting, validated the algorithm [35], and implemented the same spectral fitting equations as Nachabe et al [36, 37]. However, contrary to Nachabe et al, our algorithm does not have limitations of diffusion theory.

We emphasize that ${f_{{\text{bile}}}}$ may contain traces of bilirubin found in blood at ranges between 5.8% and 40% in humans, especially considering that bilirubin is a bile chromophore. Bilirubin blood serum levels can vary between 10 and 50 mg dl⁻¹ [38] and hemoglobin blood levels vary between 121 and 172 mg dl⁻¹ [34]. With that in mind, ${f_{{\text{bile}}}}$would represent the bile concentration added to a contribution of bilirubin varying between 5.8% and 40% of THb. The use of bilirubin absorption spectrum in our fitting is limited by the narrow wavelength range reported in the literature [39], which is not sufficient to fit a spectrum between 450 and 1590 nm.

In order to evaluate whether changes in parameters extracted from the spectral fitting were statistically significant, we used the Wilcoxon rank sum test. Levels of statistical significance were shown as start marks in boxplots for each parameter: * for p < 0.05 (statistically significant), ** for p < 0.01, and *** for p < 0.001 (statistically highly significant). Even though Wilcoxon rank sum test was used for uniform representation of p-values under the same statistical test, we are aware that two-sample t-test may be more appropriate when distribution is considered normal. Then, we performed the Anderson–Darling and Lilliefors normality tests and show the result of a two-sample t-test for the hypothesis that reflectance measurements of each organ at BT and CT have unequal means. The result of all normality and statistical tests is shown in the supplementary material.

In this study, we assessed the effect of cooling on tissue microvascularity (total hemoglobin percentage THb, oxygen saturation StO₂, average vessel diameter R, methemoglobin percentage f_MetHb), hydration (f_water), lipid content (f_lipid), bile concentration (f_bile), and microstructure (scattering amplitude α', Mie scattering power b_Mie and fraction of Rayleigh scattering f_Ray). The success of this assessment can be confirmed by expected trends of several parameters including (a) the increase of StO₂ at CT due to metabolism reduction and less oxygen consumption, (b) decrease in average vessel diameter due to vasoconstriction and (c) bile production in liver during the cooling process, as liver metabolism is still active until it reaches low temperatures. With this assessment, we have demonstrated that DRS could potentially be used for multiparametric evaluation of organ function [30, 31, 42–47], and hypothermia effects associated with tolerated cold storage times. This evaluation could be applied at stages of the transplantation procedure where intervention may be required.

Although previous studies have used DRS and non-contact optical methods retrieving similar reflectance quantities to determine tissue THb, THC, StO₂, f_water and f_lipid, studies evaluating the DRS extended wavelength range between 450 and 1590 nm such as the present study are scarce. The importance of considering the extended NIR wavelength range in the DRS spectral fitting algorithm has been demonstrated by Nachabe et al [37] through calculation of similar biochemical and microstructural parameters obtained in this study by using a diffusion-theory-based spectral fitting algorithm for wavelength ranges between 500–1000 nm and 500–1600 nm. Even though microvascular parameters (THb, THC, StO₂) and b_Mie were not significantly affected by the wavelength range extension, f_water, f_lipid, α', and Mie-to-Rayleigh fraction of the scattering (i.e. 1 − f_Ray) experienced a significant change that would not be possible to consider without the extension. Hence, extracting biochemical and microstructural parameters by using the wavelength range between 450 and 1590 nm in this study allows a more accurate calculation of f_water,more accurate calculation f_lipid,more accurate calculation α' and f_Ray.

It is important to remember that this study was based on a MC LUT-based spectral fitting algorithm instead of a diffusion-theory-based algorithm and, thus, does not have diffusion-theory restrictions (including invalidity at highly anisotropic media, and when the reduced scattering coefficient $\mu {\prime _{\textrm{s}}}$ is not substantially higher than the absorption coefficient ${\mu _{\text{a}}}$) existent at spectral regions of high absorption by blood and water. With this in mind, the wavelength range extension may affect THb, THC and StO₂. Even though no significant statistical differences were shown previously [36, 37], the absence of these differences may happen merely due to the lack of diffusion theory validity at spectral regions used to estimate microvascular parameters. Then, the determination of microstructural and biochemical parameters may be more accurate in our study, as these parameters are interdependent during their calculation by the DRS spectral fitting algorithm. The accuracy of this determination can be important to appropriately characterize heart, kidney, liver and lung tissues, which are commonly transplanted. Furthermore, understanding the effect of cooling in these organs can be relevant for several applications discussed in the next sections of this manuscript.

The combination of pathophysiological processes responsible for organ injuries is called ischemia/reperfusion injury (IRI). IRI can occur at different stages of transplantation process such as organ retrieval from the donor, organ preservation, and reimplantation in the recipient (reperfusion). However, alterations on tissue microstructure and biochemistry related to cooling alone have been poorly investigated, especially because the success of organ preservation through cold storage has been shown to depend on many variables including the type of death of the organ donor, vascular flushing organs with OPS of different compositions, cold storage times, and adoption of static or machine perfusion strategies [4, 5].

Although previous research has empirically investigated the contribution of these variables to the success of the transplantation procedure and organ preservation has been shown to be a major contributor to graft survival after revascularization [5], it is still unclear whether organ ischemia is caused by the cooling process or the actual cold storage. With this in mind, understanding the effect of cooling in the preservation of each type of organ is essential to determine whether organ preservation strategies based on cold storage are more effective than midthermic and normothermic strategies. Even though cooling has well known harmful consequences on tissues due to oxidative stress and inflammation [5] for relatively long cold exposure times, cold storage preservation techniques are still the most widespread in transplantation procedures [4, 5].

Using hypothermia to maintain organ viability was considered as one of the first strategies to contain repercussions of ischemia during organ transfer from donor to recipient [48]. Cooling is the first organism response against hypoxia, as metabolism reduction decreases the need for oxygen used in the aerobic respiration. Depression rate of metabolic reactions with cooling was quantified by using the Q10 rate created by Heilbrunn et al [49]. This rate indicates the drop of reaction rate for every 10 °C reduction in temperature. Zimmerman et al [50] and Fuhrman et al [51] showed that Q10 is approximately 2 from 20 °C to 40 °C for a number of reactions, and increases to between 3 and 5 for temperature ranges from 0 °C to 20 °C. Therefore, decreasing an organism temperature from 40 °C to 0 °C would reduce the metabolic reaction rate by 36–100 times, leading to structural and biochemical changes at subcellular level [52]. Structural changes happening within hours of cooling [52] include precipitation of F-actin microfilaments and depolymerization of microtubules upon storage of hepatocytes at 4 °C [53], while kidney cells experience microtubule disruption by cold [54]. In both liver [55] and kidney [56], mitochondria swelling and vacuolization was also observed. The aforementioned structural changes would be detectable by studies evaluating organs preserved using hypothermia, as storage may take 12–24 h. In this study, both kidney and liver tissue have already experienced microstructural changes leading to alterations in optical scattering (scattering amplitude and fraction of Rayleigh scattering) within 3 h.

Tissue/cellular biochemical changes due to cooling include inflammation (cytokine production) and oxidative stress (production of reactive oxygen species), cell edema, increased anaerobic glycolysis, cell membrane depolarization, disruption of the membrane permeability, breakdown of ion homeostasis, degradation of adenosine, accumulation of calcium, sodium, water, hypoxanthine and xanthine oxidase within the cell [5]. These changes lead to a sequence of membrane and intracellular events which culminate in protein degradation and membrane injury responsible for lethal cell injury [57], necrosis and apoptosis [58].

In this study, tissue metabolism could be assessed through parameters such as StO₂, f_bile, scattering amplitude (α'), b_Mie, and f_Ray shown in table 1. Based on these parameters, tissue oxygenation was higher for all organs at low temperatures, which suggests reduction of tissue metabolism and oxygen consumption in conditions where oxyhemoglobin is still present 3 h after animal death. In addition, α' decreases for heart, kidney and lung, b_Mie increases for liver and f_Ray decreases for heart, kidney and liver. In kidney, disruption of microtubules and membrane permeability/polarization (along with respective membrane structural changes) may be associated with the f_Ray decrease, as absence of microtubules (typical diameter of about 25 nanometers) reduces the amount of Rayleigh scattering and membrane depolarization may be accompanied by reduction of refractive index mismatches (plasma membrane thickness ranges from 5 to 10 nm). While the same changes may occur due to membrane depolarization in liver and heart, liver f_Ray reduction may be also related to F-actin microfilament precipitation and microtubules depolymerization. The mitochondria swelling and vacuolization of both kidney and liver may affect all scattering parameters α', b_Mie, and f_Ray due to the generation of Mie scattering which is the predominant type of scattering in biological tissues. In this case, larger mitochondria would generate (a) higher α' due to larger amount of refractive index mismatches within the same cell volume, (b) lower b_Mie characteristic from larger scattering particles, and (c) unknown effect on f_Ray, as more Mie scattering may happen depending on the increase of mitochondrial size while the generation of Rayleigh scattering increases due to a larger number of membranes (5–10 nm) within mitochondria. Finally, the increased f_bile in liver shows that the liver tissue was viable during the entire period of our experiment. We cannot draw any conclusion about the bile production at CTs in this study, as temporal measurements were not considered. Still, we show that bile was produced by liver from the time of mice euthanasia to the end of the cooling process (as temperatures varied from a maximum of 37 °C to a minimum of 0 °C), even though it is well known that liver produce little bile during hypothermia (less than 0.1 mg per 50 h at organ constant temperatures of 4 °C) [41].

The oxidative stress produced by reactive oxygen species are hypothesized to be one of the factors involved in the microvascular and parenchymal cell injury during organ preservation [5]. According to Guibert et al [5], 'oxygen radicals increase microvascular permeability by creating large leakage sites predominantly in the small venules'. Parks et al [59] showed that these radicals are indirectly associated with circulatory shock, which means oxygen stress will have an effect on aerobic cellular respiration and related metabolic pathways including those discussed on the previous section of this manuscript. Metabolic reactions related to aerobic cellular respiration affect the cellular levels of adenosine triphosphate (ATP) during ischemia, while those associated to increased anaerobic glycolysis generate intracellular acidosis and changes in the mitochondrial membrane permeability [5]. A certain threshold of this permeability and ATP levels determines events of apoptosis or necrosis. The amount of these events decides whether organ dysfunction will occur after organ preservation at CTs.

When cold preservation is associated to warm re-oxygenation (reperfusion with blood at BT), liver cell injury causes dysfunction in the bile secretion, drug metabolism, mitochondrial function, cell volume [60]. In general, mitochondrial permeability transition associated with mitochondrial swelling is expected from most cells [5]. However, liver endothelial cells exhibited mitochondrial ultracondensation [61] which may be involved with apoptosis [62]. Upon rewarming to 37 °C after 2 h of cold incubation in Univeristy of Wisconsin (UW) solution, Kerkweg et al [62] showed that mitochondrial alterations were fully reversible and cells remained viable, while mitochondrial fragmentation was experienced by liver cells rewarmed after 18 h of cold incubation. Biochemical and microstructural changes associated with cold-induced injury were not well determined for other organs and still depend on the combinations of (a) cooling processes used, (b) OPS and (c) oxygen supply during preservation [5, 63]. These changes determine the tolerated periods of cold ischemia, which typically ranges between 6 and 24 h and is highly dependent on the organ (6 h for heart, 8 h for lung, 12–15 h for liver, 24 h for kidney) [5].

In the present study, we have shown that, upon cooling, THb was higher for heart, StO₂ was higher for all organs, and the average vessel diameter (R) was smaller for heart, kidney and liver. Higher heart THb may be explained by the increase of vascular permeability due to the increase of oxygen radicals, as described by Guibert et al [5]. Once this permeability increases, leakage may happen on small vessels, which increases the amount of blood available in the tissue surface. Since the tissue surface is where most of the signal of our DRS probe is collected, an apparent rise in THb is sensed. In fact, a consistent increasing trend (on average) was observed for all organs, although this increase is not statistically significant. The leakage can be even more pronounced if we consider that statistically significant vasoconstriction is present in most organs (lower R values) and that small venules are more likely to leak [5]. A significant difference in THb may have been observed only in heart because it is the organ containing larger amount of blood vessels. Decreasing R is expected due to vasoconstriction at lower temperatures. Even though lung did not exhibit significant statistical difference in R, a decreasing trend was observed on average. Finally, we believe that hemoglobin molecules could still bind to oxygen to form HbO₂ throughout the duration of our DRS measurements, as StO₂ was higher for all organs upon reduction of the tissue temperature and, thus, its metabolism and oxygen consumption. We should emphasize that DRS does not only measure hemoglobin inside the blood vessels and red blood cells, but can also include free hemoglobin which may leak to tissues due to increase vascular permeability. Intra- or extracellular hemoglobin found closer to the tissue surface may bind to oxygen at higher rates. Therefore, the significant increase in StO₂ for all organs may be explained by summing up increased formation of HbO₂ and reduced oxygen consumption by tissues.

Currently, tolerated times of cold ischemia are determined by empirical tests based on the outcome of conventional transplantation protocols [4, 5]. However, this outcome depends on many factors unrelated to the organ preservation such as organ donor demographics and available resources and there is no consensus on the best organ preservation strategy [5]. Since these strategies involve hypothermia in most of cases, understanding the effect of cooling alone is essential to build tools for real-time monitoring of organ function, which can be used to determine personalized tolerated cold ischemia times for each transplantation case. For instance, one can monitor tissue blood oxygenation prior to organ preservation and after reperfusion, but still not be able change negative outcomes of the transplantation procedure, as interventions to make the procedure successful should happen at organ preservation stage. These interventions include tailoring the procedure to accommodate the appropriate optimum cold storage period once this period is accurately determined by supporting medical devices.

Although tolerated times of cold ischemia are fixed by protocols, it is clear that optimum times may differ for donation after circulatory death (DCD) and donation after brain death (DBD), as DCD cases involve organs that experience cessation of blood flow for extended periods and, therefore, earlier lack of oxygenation. These times and their negative impact in kidney transplantation before machine perfusion preservation are discussed in detail by Paloyo et al [64], who showed that DCD transplants had higher (30.5%) overall incidence of delayed graft function compared to DBD transplants (7.3%). In addition, unfavorable effects were observed for SCS higher than 6 h for DCD recipients and hypothermic machine perfusion (HMP) times higher than 36 h for both DBD and DCD recipients. Yet, although detrimental hypothermic effects have been shown through graft survival in Kaplan–Meier curves, optimum SCS and HMP times could be specified.

By understanding the effect of cold ischemia, calculation of the optimum cold storage time in both DBD and DCD transplantation cases could be based on the parameters evaluated during real-time organ function monitoring. For example, by knowing the oxygen saturation (StO₂) of each organ after circulatory death and their oxygen consumption between the interval between circulatory arrest and declaration of death, it may be possible to calculate how much tolerated cold storage time could be deduced for higher chances of transplantation success.

Our results suggest that, even though the oxygen is still consumed at body and cold storage temperatures, there may be an oxygen diffusion rate to hemoglobin inside or outside blood vessels which allows blood to be more oxygenated upon cold storage. With this in mind, StO₂ information combined with determination of oxygen inflow into tissues could be used to calculate the oxygen consumption of each organ. Furthermore, we have demonstrated that DRS could be used to measure StO₂ after cold storage in clinical settings for more precise determination of StO₂ in diverse cases. It is important to note that StO₂ can be combined with other parameters to evaluate organ viability such as f_bile for liver.

In general, OPS were developed to circumvent problems generated by vascular flush perfusion of cooled heparinized blood or diluted blood such as early vascular stasis upon graft reimplantation [4, 5, 48]. That way a replacement for blood solutions was sought by understanding the requirements of animal systems to adapt to cold exposure including (a) metabolic suppression, (b) metabolic pathways allowing minimal long-term energy supply and (c) improved defense mechanisms to revert the metabolism back to normal activity after hibernation [52]. While OPS can be designed to induce similar changes into human organs for survival during cold storage, OPS should ideally be synthetic solutions that could be available at short notice, reliably produced, and sterilized. With this in mind, Collins et al [65] designed one of the first OPS to resemble the intracellular electrolyte balance of the mammalian cells and, subsequently, other OPS were created with diverse compositions of electrolytes, buffers, impermeants and pharmacological agents [4, 5].

Ultimately, current OPS design would primarily consider empirical laboratory tests in cells or excised tissues, preclinical animal experiments, and outcome of clinical trials. Cell/excised tissue studies are typically centered on understanding the biochemical effect of OPS with a delay in analysis, whereas animal and human experiments are focused on the outcome of the transplantation. However, studies monitoring tissue biochemistry, especially in the cold storage phase are scarce, and novel OPS design may benefit from understanding biochemical and microstructural changes due to cold organ preservation at tissue level in real-time. Our study has shown that monitoring these changes with DRS is feasible and could potentially be applied in other preclinical/clinical settings. Based on our findings, metabolic changes can be evaluated during cold storage and, thus, results of metabolic modulation using OPS could be assessed using the same parameters associated with organ viability such as StO₂ and f_bile. Still, it is important to remember that changes in other parameters reported in this studied should be considered in that assessment, as variables extracted by our spectral fitting algorithm are interdependent.

With the development of the machine perfusion strategies, it is increasingly important to monitor biochemical parameters during organ preservation through hypothermic/midthermic/subnormothermic perfusion. Evaluated parameters should be related to processes improving perfusion effectiveness such as (a) prevention vasculature compression due to expansion of the interstitial space, (b) counteracting hypothermic cellular edema, and (c) in the case of oxygenated solutions, supply of appropriate gas atmosphere to reduce oxidative metabolism. Points (a) and (b) could potentially be monitored by f_water. Microvascular parameters would be challenging to monitor based on hemoglobin parameters of DRS, as blood will be replaced by OPS in perfused organs. Yet, the tissue average vessel diameter may be monitored if a chromophore with known absorption spectrum in the wavelength range probed by DRS is present in the OPS composition. Finally, DRS could be used to evaluate other OPS characteristics relevant to the success of graft survival such as the viscosity and colloid osmotic pressure (COP). COP counterbalances the filtration pressure in the bloodstream and equilibrates the water distribution between the interstitial space and vasculature. Therefore, it controls the intracellular and interstitial water content during hypothermic perfusion and postperfusion effects on blood flow, bile production and tissue injury [4]. Since COP effects influence f_water, f_bile and potentially microvasculature parameters, DRS could be a potential tool used to monitoring those effects on organ grafts during HMP. In addition, viscosity may affect organ viability due to different diffusion rates of OPS components, while at times the viability is indirectly influenced by biological processes associated with OPS composition (such as added colloids and pharmacological agents). van der Plaats et al [66] showed that UW preservation solution containing hydroxyethyl starch caused red blood cell aggregation which affected organ washout. In this case, using lower viscosity meant lower vascular resistance and improved organ viability compared to organs flushed with cold UW solution (higher viscosity). However, even though OPS viscosity is strongly affected by the temperature and COP, the temperature dependence of the viscosity of different OPS formulations as well as resulting effects on transplantation has not been thoroughly investigated [4]. The knowledge gap that needs to be investigated presents a good opportunity for optimization of OPS formulation based on biochemical parameters extracted by optical techniques combined with other analysis methods. The flowchart shown in figure 6 summarizes the potential value of DRS parameters for transplantation procedures and optimization of OPS formulations.

Zoom In Zoom Out Reset image size

The animal model used in this study focused on evaluating the effect of temperature alone on tissue microstructure and biochemistry. In order to do this, we attempted to suppress the physiological changes induced by OPS constituents, and by exposure to the environment by keeping organs interconnected in the mice body. The model we used was not intended to resemble cold storage in a real clinical scenario because this storage can be changed by multiple variables which are the reason why no consensus on the optimal strategy for SCS has been achieved up to date. With this in mind, we provided tissue biochemical and microstructural parameters which constitute a baseline for the addition of physiological effects due to other factors such as OPS of different compositions, machine perfusion, and storage at different temperatures. To properly consider the effects of such baseline, we recommend taking into account the following differences between cooling in our animal model and that performed in a clinical setting:

• (a)  
  Temperature distribution and heat conduction: in organ donation and preservation with cold storage methods, the first cooling process (prior to storage) happens due to perfusion of low-temperature OPSs, which cool organs starting from perfused tissue volumes reached by OPS within them (heat exchange from the inside to outside the organ). This cooling process is different from the homogeneous cooling performed in this study, as heat is exchanged from outside the organ to its inside. Due to differences in temperature distribution, changes in scattering properties during the cooling process will depend on the microstructure of each organ, as cooling from inside or outside the organs may shrink complex tissue structures in many different ways (e.g. position and thickness of tissue layers) and lead to different distributions of scattering particles over the tissue volume. In addition, tissue scattering is also expected to be affected by the state of lipid molecules which may transition from a 'liquid' to a 'solid' phase upon temperature reduction from BTs to cold storage temperatures. In this case, cooling from outside the organ might change the scattering locally at the tissue surface compared to cooling from the perfused locations, as cellular and subcellular structures containing lipids may have their scattering properties altered. However, it is worth noting that we would not expect variations in the tissue lipid content due to differences in temperature distribution within the organ during relatively short cooling procedures.
• (b)  
  Cooling interval: Perfusing the organ with OPS cools organs faster than cooling organs by placing them in a cold environment, as perfusion is constantly replacing the OPS volumes which exchanged heat with tissue (higher OPS temperatures after heat exchange) by new volumes with low-temperature OPS. Thus, the cooling interval from BT to CT may be reduced in medical settings. We expect faster cooling to be associated with quicker stabilization of StO₂ (dependent on oxygen consumption) and water content (f_water; dependent on organ dehydration rates). With this in mind, absolute StO₂ and f_water values obtained in our study are not relevant to the reality of transplantation procedures, as short-term StO₂ and f_water readings in the real case scenario will depend on the time interval taken for organ removal and transportation prior to be perfused with OPS. Yet, we believe that the StO₂ and f_water variation investigated in this study are important to be considered.
• (c)  
  Organs interconnected in the body: The variation of StO₂ is not completely separated for each organ if there is interaction between organs due to connection molecular diffusion within blood vessels. However, we believe that this interaction is limited, as there is no blood flow or circulation in our animal model, and the molecular diffusion is expected to be relatively slow to cause StO₂ differences during the cooling process.
• (d)  
  Organ viability: The impact of factors described above on organ viability is still unknown. The short-term physiological effects induced by these factors during cooling may affect organ viability and, thus, impact viability-associated parameters such as f_bile in liver tissue.

Finally, the probed depth of DRS in the tissues targeted in this study is estimated to be up to 1 mm slightly depending on the wavelength. The probed depth can be increased by using fiber optic probes with larger source-to-detector distances or by using time-resolved techniques. We would like to emphasize that our study provides only a proof-of-principle for the DRS technology to be deployed in organ preservation monitoring and characterizes the effect of temperature alone during cold preservation of mouse organs, which can be further translated into technology involving human organs in future research.