Current Insights into the Metabolome during Hypothermic Kidney Perfusion—A Scoping Review

This scoping review summarizes what is known about kidney metabolism during hypothermic perfusion preservation. Papers studying kidney metabolism during hypothermic (<12 °C) perfusion were identified (PubMed, Embase, Web of Science, Cochrane). Out of 14,335 initially identified records, 52 were included [dog (26/52), rabbit (2/52), pig (20/52), human (7/52)]. These were published between 1970–2023, partially explaining study heterogeneity. There is a considerable risk of bias in the reported studies. Studies used different perfusates, oxygenation levels, kidney injury levels, and devices and reported on perfusate and tissue metabolites. In 11 papers, (non)radioactively labeled metabolites (tracers) were used to study metabolic pathways. Together these studies show that kidneys are metabolically active during hypothermic perfusion, regardless of the perfusion setting. Although tracers give us more insight into active metabolic pathways, kidney metabolism during hypothermic perfusion is incompletely understood. Metabolism is influenced by perfusate composition, oxygenation levels, and likely also by pre-existing ischemic injury. In the modern era, with increasing donations after circulatory death and the emergence of hypothermic oxygenated perfusion, the focus should be on understanding metabolic perturbations caused by pre-existing injury levels and the effect of perfusate oxygen levels. The use of tracers is indispensable to understanding the kidney’s metabolism during perfusion, given the complexity of interactions between different metabolites.


Introduction
Hypothermic perfusion preservation, also called hypothermic machine perfusion, reduces the risk of delayed graft function [1]. Studies also suggest improved graft survival in kidneys donated after brain death (DBD), though this is not the case for kidneys donated after circulatory death (DCD) [2][3][4]. A recent randomized controlled trial suggests that actively oxygenating the perfusate during hypothermic perfusion of older DCD kidneys improves kidney function and survival [5]. The underlying mechanisms by which hypothermic perfusion exerts its effect are still incompletely understood.
To understand the mechanisms that drive the effect of hypothermic perfusion preservation and how this technique might be further improved, it is essential to understand "on-pump" kidney behavior. A key concept of hypothermic preservation is that metabolism, and therefore cellular metabolic requirements, are minimized. However, although the metabolic rate below 4 • C is reported to be about 5-10% of that at body temperature [6,7], there is still active metabolism in the cold. Furthermore, the preservation temperature during hypothermic perfusion often does not reach below 4 • C. The metabolic activity in detection), confounding, study power, the strength of causality in the association between interventions and outcomes, and other factors [14].

Results
A systematic search of online databases, performed on 2 February 2023, resulted in the identification of 14,335 records. After duplicate removal, 9794 records remained, of which 9596 were excluded based upon predefined inclusion and exclusion criteria at the time of initial screening and 152 at the time of full-text screening, leaving 46 included articles. From the reference lists, another 1507 potential papers were identified leading to six additional inclusions. In total, 52 papers were included in this scoping review. Figure S1 shows the flowchart. The full data extraction table can be accessed online [12]. Included articles were published between 1970 and 2023 (Table S3).

Quality and Risk of Bias Assessment
For animal studies, the risk of bias was most often "unclear" or "high" because essential information was often not reported ( Figure S2, Table S4). For human studies, the overall quality was better, most often "good" or "fair" ( Figure S2, Table S5).

Hypothermic Perfusion Set-up
The majority of articles reported on animal experiments (48/52) in dogs (26/52), rabbits (2/52), and pigs (20/52) ( Figure 1). Seven studies reported on the perfusion of human kidneys (7/52), of which three were transplant studies [8,15,16]. Some papers report on both animal and human kidney perfusions (3/51) (Table S3). Kidneys were exposed to variable ischemic injury, induced by introducing warm ischemia (5 to 240 min; clamping of the renal artery (and vein) before procurement or procurement after death) or exposure to cold storage before perfusion (up to 20 h) (Table S3). After hypothermic perfusion, kidneys were transplanted in 22 studies [8, or re-perfused with a blood-based perfusate in one study [36]. Kidneys were flushed with different solutions (detailed in the extraction table [12]) before mounting on the perfusion device. to increase transparency and applicability of results [13]. For studies with human organs, methodological quality was assessed using the National Institutes of Health (NIH) scoring tools. These tools include items for evaluating potential flaws in study methods or implementation, including sources of bias (e.g., patient selection, performance, attrition, and detection), confounding, study power, the strength of causality in the association between interventions and outcomes, and other factors [14].

Results
A systematic search of online databases, performed on 2 February 2023, resulted in the identification of 14,335 records. After duplicate removal, 9794 records remained, of which 9596 were excluded based upon predefined inclusion and exclusion criteria at the time of initial screening and 152 at the time of full-text screening, leaving 46 included articles. From the reference lists, another 1507 potential papers were identified leading to six additional inclusions. In total, 52 papers were included in this scoping review. Figure S1 shows the flowchart. The full data extraction table can be accessed online [12]. Included articles were published between 1970 and 2023 (Table S3).

Quality and Risk of Bias Assessment
For animal studies, the risk of bias was most often "unclear" or "high" because essential information was often not reported ( Figure S2, Table S4). For human studies, the overall quality was better, most often "good" or "fair" ( Figure S2, Table S5).

Hypothermic Perfusion Set-up
The majority of articles reported on animal experiments (48/52) in dogs (26/52), rabbits (2/52), and pigs (20/52) (Figure 1). Seven studies reported on the perfusion of human kidneys (7/52), of which three were transplant studies [8,15,16]. Some papers report on both animal and human kidney perfusions (3/51) (Table S3). Kidneys were exposed to variable ischemic injury, induced by introducing warm ischemia (5 to 240 min; clamping of the renal artery (and vein) before procurement or procurement after death) or exposure to cold storage before perfusion (up to 20 h) (Table S3). After hypothermic perfusion, kidneys were transplanted in 22 studies [8, or re-perfused with a blood-based perfusate in one study [36]. Kidneys were flushed with different solutions (detailed in the extraction table [12]) before mounting on the perfusion device. Several perfusion devices were used, from homemade to commercially available devices (Table S3). The latter were sometimes adjusted to fit the aim of the research. Common features of these perfusion circuits were a reservoir and tubing, and a pump to circulate the perfusate. Cooling (2-12 °C) was accomplished by a heat exchanger, ice surrounding the reservoir, or a combination of both. The majority of the circuits used pulsatile perfusion (roller, centrifugal, or peristaltic pump); continuous non-pulsatile perfusion was used in two studies [37,38]. In most studies, perfusion was pressure controlled (25 to 60 mmHg) with a maximal pressure of 30 mmHg in recent studies. Several perfusion devices were used, from homemade to commercially available devices (Table S3). The latter were sometimes adjusted to fit the aim of the research. Common features of these perfusion circuits were a reservoir and tubing, and a pump to circulate the perfusate. Cooling (2-12 • C) was accomplished by a heat exchanger, ice surrounding the reservoir, or a combination of both. The majority of the circuits used pulsatile perfusion (roller, centrifugal, or peristaltic pump); continuous non-pulsatile perfusion was used in two studies [37,38]. In most studies, perfusion was pressure controlled (25 to 60 mmHg) with a maximal pressure of 30 mmHg in recent studies. Acellular perfusates were used without the use of an oxygen carrier. Between 1970 and 1986, these were plasma-or albumin-based and often a vasodilator, heparin, antibiotics, corticosteroids, allopurinol, insulin, and buffers to maintain pH were added ( Table S6). As of 1980, synthetic perfusates were introduced. Albumin was replaced by synthetic oncotic products (mannitol, hydroxyethyl starch, etc.) and the addition of less permeable anions (like gluconate and lactobionate) prevented further hypothermic-induced cell swelling in contrast with other anions (like chloride) [39]. Adaptations were made to include adenine nucleotide derivates and electrolytes were added in extracellular or intracellular concentrations. After 1986, all studies were performed with synthetic solutions, mostly formulations of Belzer's Machine Perfusion Solution (MPS) [39]. In some studies, carbohydrates, amino acids, or fatty acids were added to the perfusate (Table S6).
When ischemically injured kidneys were perfused with an oxygenated glucose-containing perfusate, glucose concentrations decreased with increasing levels of lactate and pyruvate [25].

Fatty Acid Metabolism
Perfusate-free fatty acids [29] and lipids [63] decreased during the oxygenated perfusion of minimally injured kidneys. The decrease in perfusate triglycerides and tissue phospholipids was less pronounced when oleate (a long-chain fatty acid) was administered during perfusion, with a more pronounced decrease in perfusate neutral lipids but high retention of tissue triglycerides and neutral lipids [63].

Energy Metabolism
In minimally injured kidneys, ATP levels increased during oxygenated perfusion while in injured kidneys, tissue ATP levels decreased rapidly during warm ischemia, with restoration during oxygenated perfusion [56]. Nevertheless, Pegg et al. did not observe adenine nucleotide restoration [25] and Kahng et al. described a variable range for each adenine nucleotide and a suboptimal energy charge during perfusion [57].

Carbohydrate Metabolism
A decrease in perfusate glucose and an increase in perfusate lactate concentration was seen during oxygenated perfusion [27,42,[46][47][48]. Only one study showed no change in perfusate glucose level with a low lactate and non-measurable pyruvate during oxygenated perfusion with bovine serum albumin [34]. This glucose drop was more pronounced when glucose was a component of the perfusate [27] and less pronounced when amino acids [42] or fatty acids [46,47] were added. Interestingly, Slaattelid et al. also found a different perfusate glucose uptake and release pattern when glucose was administered intravenously before kidney procurement [27]. Three studies found a decrease in tissue glucose regardless of whether the perfusion medium was glucose-rich or -poor [27,37,38]. Tracer studies showed active glucose metabolism, with the detection of 14 C-CO 2 and 14 Clactate [27,[46][47][48]. An increase in the lactate/pyruvate ratio was seen in the studies by Grundmann et al. [23] and Pettersson et al. [48]. Tissue lactate showed no change [38] and was the highest metabolite measured [57], even increasing after 72 h when glucose was added [37].

Amino Acid Metabolism
Many amino acids were released during the oxygenated perfusion of minimally injured kidneys, with an increase in urea and ammonia [42]. The increase was most pronounced for alanine and taurine [41]. No increase was seen for glutamine, proline, aspartate, and cystine [42].
In contrast, when amino acids were added to the perfusate, a decrease in many amino acids (glutamine, proline, glycine, aspartate, arginine, (iso)leucine, methionine, valine, and serine) was seen with an increase only in threonine, tyrosine, ornithine, taurine, and alanine levels [41,42]. Labeled (cyclo)leucine and threonine were incorporated into proteins; more so for leucine which is also reflected in the higher label recovery in CO 2 from leucine than threonine [41].
In tissue, phospholipids decreased with fatty acid-free perfusate [47]. When a fatty acid-rich perfusate was used, phospholipid levels did not change [49] or decreased [48]. Tissue cholesterol levels remained unchanged in three studies in which kidneys were pumped for 2 to 6 days [47,49,53] and decreased in one study after 6 days of perfusion [48].

Metabolism of High-Energy Molecules
In minimally injured kidneys, ATP levels and total adenine nucleotide levels decreased during oxygenated perfusion but the energy charge potential was optimal [37]. While in ischemically injured kidneys, ATP levels increased to near normal after 24 h of oxygenated perfusion but total adenine nucleotide levels remained the same [28] or had a variable range with suboptimal energy charge [57].
In two studies, non-radioactively labeled 13 C-glucose [60,64] was used to further clarify the metabolism. Details are listed in Tables 3 and S9.

Carbohydrate Metabolism
Perfusion with oxygenated glucose-free perfusate led to low and decreasing levels of tissue glucose and very low and stable lactate levels in minimally injured kidneys [37].
In ischemically injured kidneys, perfused without active oxygenation, glucose changes seem dependent on the glucose concentrations in the perfusate. Indeed, no changes in glucose levels were seen during long-term perfusion with MPS (72 h perfusion with perfusate changes every 24 h) [59]. In a similar 72 h experiment with 24 h renewals of UHK solution, containing less glucose and mannitol compared to MPS, perfusate glucose levels dropped [59]. Similarly, other studies investigating short-and long-term non-oxygenated perfusions of pig kidneys with MPS found no statistical change in perfusate glucose levels [8,16,58] or an increase in glucose [15] with increases in perfusate and tissue lactate [8,15,16]. A similar study with human kidneys showed a glucose and lactate increase in the perfusate [15]. Only one group studied glucose changes during the oxygenated perfusion of injured kidneys and found increased perfusate glucose and lactate levels with similar or increased tissue lactate [19][20][21][22]. Perfusate lactate levels were lower when high oxygen concentrations were given [20][21][22]64].
Tracer studies show active glucose metabolism during (non-)oxygenated perfusion of ischemically injured kidneys. When 13 C-glucose was infused, 13 C-lactate appeared in the perfusate and tissue [60,64]. Less labeled lactate was recovered in tissue after highly oxygenated perfusion [64].
Mannitol did not change with non-oxygenated MPS [21] and increased during oxygenated perfusion [21]. Ribose levels remained stable in two studies [8,15] and decreased in one study [16].
Oxidized glutathione in the perfusate also decreased or was undetectable [15,16]. There is some suggestion that higher oxygen concentrations increase the levels of tissue glutathione [64].
During oxygenated perfusion of minimally injured kidneys with gluconate-based perfusates, tissue phospholipids were studied. An initial decrease was seen in the first 24 h followed by an increase in phospholipids [54].

Energy Metabolism
Belzer's MPS contains adenine and ribose after studies showed higher ATP concentrations with these additives compared to adenosine during oxygenated perfusion of minimally injured kidneys [32,52,55]. Furthermore, with MPS, ATP content increases with higher oxygen concentrations [50].

Discussion
Commendable work contributing to our understanding of kidney metabolic behavior during cold perfusion has been performed over the past 50 years. Nevertheless, it remains incompletely understood.
From compiling the findings of this scoping review, and in particular those of "tracer studies" during which a (non)radioactively labeled metabolite is added to the perfusion circuit, it is clear that kidneys are metabolically active during cold perfusion. However, key pieces of the puzzle are missing. The fact that metabolism is intrinsically a complex network of interacting biochemical reactions that can be influenced by numerous factors complicates the interpretation of findings. Indeed, perfusate composition, oxygenation, pre-existing kidney injury, and perhaps even pre-donation nutrient availability seem to influence metabolism during hypothermic perfusion preservation. Furthermore, key enzymes are likely to be influenced by the low temperatures and this in turn will influence metabolism.
Tracer studies have shown the oxidation of glucose, amino acids, and fatty acids to CO 2 [27,41,42,[46][47][48][49] but it is unclear if and how oxygenation levels and pre-existing injury affect this. A few older studies suggest gluconeogenesis from fatty acids [47,48] and the incorporation of amino acids into proteins when these are provided [41]. Indeed, kidney metabolism seems dependent on the perfusate composition. Administration of carbohydrates, amino acids, and lipids seems to change the uptake and release patterns of these metabolites. In almost all studies, glucose was added to the perfusate as an energy source. Interestingly, glucose and lactate metabolism changed when other substrates such as amino acids or fatty acids were added to the perfusate, suggesting these metabolites could have competing interests as energy sources. In vivo, fatty acids and amino acids can feed into the TCA cycle either as acetyl-coenzyme A or another TCA-cycle intermediate and serve as important energy sources [67]. In physiological, in vivo conditions, the kidneys play a role in the synthesis and inter-organ exchange of amino acids [68]. It is unclear how the absence of inter-organ exchange influences amino acid metabolism during cold perfusion. Furthermore, the in vivo renal metabolism of alanine, (iso)leucine, and valine changes during fasting and feeding states [68,69] and this might influence metabolism during cold perfusion as well.
The oxygenation level of the perfusate also seems to affect kidney metabolism during cold perfusion. Indeed, a reasonable number of studies showed differences in kidney metabolism when comparing different oxygen levels. Concerning lactate metabolism, there seems to be a reduced increase in lactate when the perfusate was actively oxygenated compared to non-oxygenated perfusates [21,22,40,43,64]. ATP levels and the supported energy charge are higher when kidneys are oxygenated during perfusion [20,21,36,43,45,50,[64][65][66].
Injury levels are likely to play a role as well, but this is less clear as only a few studies directly compared kidneys with different injury levels. Pre-existing warm ischemia depletes ATP levels and studies suggest variable reconstitution of ATP during oxygenated perfusion [28,30,36,43,56,65,66]. Observations of other metabolites in minimally injured kidneys are less clear.
The potential implications of these observations in a clinical setting are important. Indeed, supporting the metabolically active kidney during hypothermic perfusion seems the logical next step. Actively oxygenating the perfusate during the hypothermic perfusion of older DCD kidneys improved graft outcomes compared to standard non-oxygenated hypothermic perfusion in a recent randomized controlled trial [5]. On the other hand, there was no evidence that short (2 h) reconditioning of expanded-criteria donor kidneys, following cold storage, improved graft survival compared to cold storage alone [70]. Perhaps 2 h is too short to change metabolic behavior after a long period of cold storage or perhaps DBD organs respond differently to additional oxygen compared to DCD.
These findings need to be interpreted with a degree of caution. The majority of studies have a considerable risk of bias and studies were conducted over the course of 50 years with numerous changes in perfusate and perfusion conditions. Many studies report on experiments with dog kidneys. It is important to realize that dog kidneys have a higher tolerance to ischemia reperfusion injury, while pig kidneys are more similar to human kidneys [71,72]. Even though pig kidneys are physiologically and anatomically comparable to human kidneys, there is only limited evidence that pig kidney metabolism is the same as that of human kidneys, however, it is reassuring that Nath et al. found comparable metabolites in pig and human kidney perfusates [8].
As with all scoping reviews, it is possible that some relevant articles were not identified or that relevant studies were published after the search. We limited the chance of missing relevant articles by setting up a broad search strategy in collaboration with experienced biomedical reference librarians. Furthermore, the references of included articles were searched to identify any articles that might have been missed in the search.

Conclusions
In conclusion, kidneys are metabolically active during cold perfusion preservation. This metabolic activity is incompletely understood and is influenced by a multitude of factors including perfusate composition, oxygenation level, and likely pre-existing injury. It is clear that a greater number of well-designed (pre-)clinical studies are necessary to understand this behavior. In the modern era, with increasing DCD donations and the emergence of hypothermic oxygenated perfusion, the focus should be on understanding metabolic perturbations caused by pre-existing injury levels and on the effect of perfusate oxygen levels. The use of tracers in such studies is indispensable to understanding the metabolism, given the complexity of interactions between different metabolites.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/jcm12113613/s1, Figure S1 Flow chart of the systematic search identifying published papers reporting on kidney metabolism during hypothermic perfusion; Figure S2: Risk of bias assessment in 30 studies identified, using SYRCLEs tool [13]; Table S1: Search string in databases Pubmed, Embase, Cochrane Library and Web of Science Core Collection; Table S2: Inclusion and exclusion criteria; Table S3: Overview of study set-up and perfusion characteristics; Table S4: Detailed Risk of Bias Assessment using SYRCLE's tool for articles reporting on animal studies; Table  S5: Quality assessment of studies including human kidneys according to the NIH quality assessment score; Table S6: Composition of different perfusion solutions studied in this review; Table S7: Extended summary of studies reporting on kidney metabolism with plasma-based perfusates; Table  S8: Extended summary of studies reporting on kidney metabolism with albumin-based perfusates; Table S9: Extended summary of studies reporting on kidney metabolism with synthetic perfusates. Refs. [73,74]   Data Availability Statement: The dataset has been deposited in RDR, KU Leuven's data repository, and is publicly available via https://doi.org/10.48804/AMSYVO [12].