LDH was one of the first proteins to be studied in the renal perfusate. This cytosolic enzyme is responsible for the conversion of pyruvate to lactate by NADH under normoxic and hypoxic conditions (20). In total, five tetrameric isoenzymes consisting of LDHA and LDHB subunits have been previously described (18,21,22). They differ in their degree of substrate affinity, electrophoretic mobility, inhibition constants and tissue distribution (18). In an experimental rat model, LDHA was previously found to be mainly present in the proximal renal tubules, whilst LDHB was found to be mainly expressed in the distal part of the nephron (23).

In relation to KTx, LDH has been documented to serve as a non-specific marker of cell injury (18). Perfusate LDH activity is not applicable in the assessment of renal function prior to KTx. Despite the studies showing a positive association between LDH concentration and DGF (15,24-26), there are no randomized controlled trials that have confirmed the viability of LDH measurements for the assessment of kidney quality. Furthermore, LDH is not a kidney-specific biomarker (18,27,28). Guzzi et al (15) previously found that besides glutathione S-transferase (GST), LDH levels were the most commonly reported parameter in the renal perfusate. Elevated LDH levels have been reported to associate positively with DGF in cases of donation after cardiac death (DCDs) (24-26), but not in all cases of donations after brain death (DBDs) (29,30). According to previous studies, (15,24-26,31,32) where LDH levels were measured only in perfusate from DCDs, LDH levels were found to be positively associated with DGF in all cases. By contrast, in studies describing the measurement of LDH levels in the perfusate of kidneys from DBDs, a positive association between DGF and LDH levels was only found in two (29,30) of the four selected studies found (15). In addition, this meta-analysis (15) also showed that LDH activity should not be considered as a marker of PNF in patients with DCDs grafts. Further analysis of cases of DBDs and DCDs demonstrated that although the LDH levels were associated positively with DGF in the majority of the included studies, following multivariable analysis only one study showed a weak association (32). In another meta-analysis previously conducted by Bhangoo et al (16), perfusate LDH activity was found to be a predictor of DGF, PNF and graft failure (GF). However, this previous analysis lacked a donor-type subgroup assessment (16). A number of the included studies revealed a significant association between LDH activity and DGF, PNF (31,33). Among the different types of donors, LDH activity was the highest in kidneys from DCD cases, which reflected the greatest extent of damage prior to KTx among the studied organs (34). Perfusate assessment has been used to estimate renal cell injury after ischemia in a previous study (29), where LDH levels were previously found to be increased during the ischemic period (29). This phenomenon also associated positively with damage to renal tubular cells detected by electron microscopy. The lesions found include mitochondrial swelling and loss of detail in the inner cristae, swelling and detachment of the basolateral tubular membrane connective tissue (29).

GSTs form one of the most frequently studied groups of biomarkers obtained from the renal hypothermic machine perfusate (15,16). In humans, the GST family consists of three subpopulations according to their localization: Cytosolic, mitochondrial and microsomal (35). GSTs can be divided into subclasses, designated by Greek letters that differ in their levels of expression in humans. α-GST is abundantly expressed in liver and renal proximal tubular cells, whereas distal tubular cells mainly express the p-GST isoform (36,37). Although GSTs mainly function as catabolic agents to catalyze the detoxification of xenobiotics within glutathione, they have also been reported to mediate cell signaling. p-GST isoform can inhibit the function of kinases involved in the MAPK pathway, which regulates cell proliferation and cell death (38).

GSTs are considered one of the most valuable and reliable non-invasive markers for pre-transplant DGF prediction, but their utility in forecasting long-term outcomes, PNF and discarding kidneys from transplantation remains controversial (15,16). Significant differences were observed in the accuracy of the predictive values of different isoforms for DGF. Hall et al (39) previously found that p-GST measured at baseline and at the end of hypothermic machine perfused (HMP) kidneys from deceased donors was associated positively with DGF. The relative risk ratio of DGF for each log-unit increase at baseline and post-perfusion p-GST concentration after adjustment (for black ethnicity, male sex, history of previous kidney transplant, diabetes as the cause of end-stage renal disease, need for blood transfusion before KTx, number of human leukocyte antigen mismatches, body mass index in kg/m² and duration of dialysis before transplant) was found to be 1.14 (95% CI, 1.00-1.28) and 1.33 (95% CI, 1.02-1.72), respectively. α-GST did not associate with DGF. In addition, the concentrations of both iso-enzymes did not differ between the transplanted and discarded organs at start and just before end of perfusion, showing that GSTs levels could not be used as a tool for discarding kidneys from KTx.

Perfusate GST levels have been previously compared with HMP parameters to improve the predictive value of GST. However, the value of this prospect remains controversial, where no clear association has yet been established. In a previous study by Moers et al (33), a correlation between the studied biomarkers [such as LDH, aspartate aminotransferase, N-Acetyl-/β-glucosaminidase, GSTs, alanine aminopeptidase (Ala-AP) and heart fatty acid binding proteins (H-FABP)] and pump parameters (such as cold ischemia time and renal vascular resistance) was not confirmed. Hall et al (39) also previously found no correlation between pump parameters (such as renal resistance at 4 h of perfusion and perfusate flow at 4 h) and concentrations of GST isoforms. By contrast, Qiao et al (40) found that the model combining end-of-perfusion vascular resistance and GST [area under the curve (AUC)=0.888; 95% CI, 0.842-0.933] significantly improved DGF predictability compared to using terminal resistance (AUC=0.756; 95% CI, 0.693-0.818) or GST alone (AUC= 0.729, 95% CI, 0.591-0.806).

Perfusate GST activity has also been investigated using other potential biomarkers of DGF. Gok et al (41) reported that GST with Ala-AP activity or GST activity with FABP concentration corresponded positively to renal injury during ischemia. However, its role in predicting the occurrence of PNF and in kidney discards for transplantation remains unclear (33).

Ischemic injury during HMP leads to the release of histones from disrupted cells (42), including extracellular histones H2A, H2B, H3 and H4. Each nucleated cell contains nucleosomes built from these four elements in an equimolar manner (43). Following the induction of pathological conditions, such as ischemia, histones are released from dying cells into the extracellular space (44,45). Using an experimental porcine model of warm ischemia and machine perfusion of retrieved kidneys, extracellular histones were reported to be cytotoxic to endothelial and renal epithelial cells in a temperature-dependent manner (46). Furthermore, histone release is higher during sub normothermic machine perfusion (SNMP) at 28°C compared with hypothermic machine perfusion (HMP) at 4°C (46). Under normal conditions, histones are responsible for nucleosome assembly (47), chromatin stability (47) and epigenetic modifications in transcription, replication and DNA repair (48). However, in the extracellular space, they function as damage-associated molecular pattern proteins (DAMPs) (49).

A previous retrospective study of 309 kidneys from DCDs was conducted on extracellular histones in the renal perfusate. van Smaalen et al (42) measured histone H3 levels at 1, 2 and 4 h of HMP. In uncontrolled DCD kidneys (Maastricht categories 1 and 2), no difference in H3 levels could be found compared with the PNF, DGF and IF (immediate function) groups. In the control cohort (Maastricht categories 3 and 4), a higher H3 concentration in perfusate was associated with DGF [univariable odds ratio (OR)=3.071; 95% CI, 1.588-5.941; multivariable OR=2.433; 95% CI, 1.253-4.723]. When comparing two groups of patients with histone H3 concentrations above and below the median, Kaplan-Meier analysis revealed a significant difference in graft survival in favor of lower concentrations. In particular, the 1- and 5-year survival rates were significantly superior in the lower H3 group (at 1 year, 84 vs. 72%; at 5 years, 79 vs. 66%) (50).

Ischemia leads to the overloading of proximal tubular cells with free fatty acids (FFAs) and tubulo-interstitial damage (51,52). During oxidative stress, FFAs bind to fatty acid binding proteins (FABPs), thereby preventing their intracellular accumulation and further oxidation (53). In addition, they can regulate the expression of genes involved in FFA metabolism and promote their β-oxidation in peroxisomes (53). To date, two isoforms have been described in human proximal tubular cells, namely liver (L)-FABPs and heart (H)-FABPs, which are also present in distal tubular cells (54). Under normal conditions, FABPs are involved in the cellular uptake and utilization of FFAs from the plasma by rapid incorporation into triacylglycerols and phospholipids, promotion of FFA metabolism in mitochondria or peroxisomes, regulation of intracellular cholesterol metabolism, regulation of gene expression involved in lipid metabolism and modulation of cell proliferation (54,55).

Perfusate FABPs have been reported to associate with DGF. Moers et al (33) found that H-FABP was an independent predictor of DGF but only had a moderate prognostic value. In another previous study, L-FABP levels were reported to be associated with estimated glomerular filtration rate (eGFR) 6 months after KTx (56). To the best of our knowledge, the association between FABP levels in the renal perfusate and PNF has not been studied. Sun et al (57) could not confirm the predictive value of FABPs for DGF and 3-year post-transplant function (expressed by eGFR value) in a retrospective study of cases of DCDs. Perfusate FABPs levels have been shown to associate with high vascular resistance and early graft dysfunction (33). Fig. 1 presents a list of molecules that can be released after renal tubular cells disruption upon necrosis and apoptosis resulting from ischemia.

NGAL is a protein that is originally synthesized in the bone marrow and is stored in neutrophil granules (58). NGAL has been previously detected in the lung, colon, trachea and renal epithelium (19). NGAL is capable of binding siderophores. Therefore, it has been postulated that NGAL can serve as a bacteriostatic agent by sequestering iron (59) and as a regulator of extracellular iron-induced injury (60). Production in non-hematopoietic locations has been associated with systemic inflammation, where one such stimuli is IL-1β (61). NGAL is typically released in response to ischemia (62) and nephrotoxic agent (such as cisplatin) contact (63). In addition, ATP depletion induces NGAL mRNA expression, leading to the production of NGAL (62). In this context, necrosis and cell apoptosis is a factor that can trigger NGAL mRNA expression in renal tubular cells (63).

In a previous systematic review, Guzzi et al (15) found three studies on NGAL (29,32,56), but only one (56) showing a positive association between NGAL concentration in the HMP kidney perfusate and DGF occurrence. The study by Moser et al (29) found that NGAL levels are not significantly different between the DGF and non-DGF groups. In another study by Hoogland et al (32), NGAL concentrations measured 4 h after initiation of perfusion were not reported to be significantly associated with the risk of DGF or PNF. Furthermore, the post-HMP NGAL concentration was found to be associated with 6-month eGFR but not with PNF (56). In another study on discarded human kidneys, NGAL levels were found to be markedly higher after 6 h of normothermic machine perfusion (NMP) administration in organs from donors with higher serum creatinine levels (sCr) at the time of retrieval (64).

However, care must be taken when examining the perfusate for the presence of NGAL and interpreting the available data. NGAL is secreted in different forms, including monomeric, dimeric and conjugated with MMP-9 (65). Available assays utilizing antibodies have not been suitably adapted to measure the monomeric and heterodimeric forms released by injured or stressed tubular cells (19). Another important issue is the timing of sample collection from the perfusion machine. NGAL synthesis and release are considered to be altered by hypothermia (15). To determine the most appropriate time interval, measuring NGAL in the renal perfusate not only after 4 h, but also after 14-16 h of HMP, are of importance to confirm the effect of hypothermia and time of perfusion on its concentrations.

MMPs have been recognized for their roles in ischemia and reperfusion processes. They have been studied in the context of ischemia-reperfusion injury in the central nervous system (66), lungs (67), liver (68), heart, (69,70), skeletal muscle (71), retina (72) and kidney (73-77). MMPs, also known as matrixins, are zinc-dependent enzymes that are located mainly in the extracellular space (78). However, MMP-1, MMP-2 and MMP-11 have been found to function intracellularly. To date, 23 MMPs have been identified in humans (79). They are first synthetized as pre-proenzymes and the majority are secreted into the extracellular space. MMPs are typically produced after stimuli, such as by inflammatory cytokines (TNF-α and IL-1), free radicals, reactive oxygen species (ROS), oncogenic cellular transformation, physical stress and chemical agents (78). In addition, a number of soluble factors can induce the expression of MMPs and molecules on the cell surface, including leukocyte function-associated antigen-1, intercellular adhesion molecule-1-mediated cell adhesion very late antigen 4 and vascular cell adhesion molecule (VCAM)-1 (80). The interaction between gp39-CD40 in monocytes and α5β1 integrin-fibronectin can also lead to MMP expression (80). Secreted pre-proenzymes are activated in the extracellular space by tissue, plasma or opportunistic bacterial proteinases. In particular, the urokinase-type plasminogen activator/plasmin system has been reported to serve a role in MMP activation (81). Pro-MMP-2 has been reported to be activated on the cell surface (82). Pro-MMP-2 is localized in podosomes or the invadopodia (83,84). This activation process requires the activation of the first membrane-type MMP (MT1-MMP), and the tissue inhibitor of MMP-2 (TIMP-2)-bound MT1-MMP (78). The TIMP-2 in the latter complex binds through its C-terminal domain to the hemopexin domain of pro-MMP-2, which is assumed to localize the zymogen close to the active MT1-MMP (78). It is also activated by plasmin produced by urokinase-type plasmin activator (from an inactive form), which is attached to a high-affinity cell membrane binding site by a specific N-terminal sequence of its non-catalytic chain (82).

MMP-2 and MMP-9, also known as gelatinases, are the most frequently investigated MMPs in KTx and organ preservation. MMP-2 activity contributes to the proteolysis of collagen types I, IV, V, VII, X, XI and XIV, gelatin, elastin, fibronectin, laminin and aggrecan (69,85). By contrast, MMP-9 targets include collagen types IV, V, VII, X and XIV, gelatin, aggrecan, elastin, entactin and fibronectin (86). Data on the presence of MMPs in the renal perfusate and their predictive value remain scarce. Moser et al (29) previously investigated the activity of NGAL, LDH and MMPs in renal perfusates from DBDs and donors after controlled cardiac death (cDCDs). MMPs could be detected in all samples and a significant positive association between perfusate MMP levels and the occurrence of DGF was reported. However, this association was not observed for NGAL and LDH levels. Additionally, higher concentrations of MMP-2 and MMP-9 were observed in the cDCDs group compared with those in the DBDs group. In another previous proteomic analysis on three types of donors, namely living kidney donors (LKDs), DBDs and cDCDs, the levels of MMPs were found to be ranked in the following manner: LKDs < DBDs < cDCDs (34).

Previous studies using animal models have shown that different storage methods may influence MMP secretion. In one such study, it was observed that HMP reduced MMP-9/NF-κB-dependent expression compared with static cold storage (SCS) in a rabbit model (87). Notably, HMP did not affect MMP-2 expression. In another study performed by Sulikowski et al (88) on a rat model, the authors investigated the impact of University of Wisconsin solution (UW) and EuroCollins solution (EC) on the gene expression of MMP-2 and tissue inhibitor of MMP-2 (TIMP-2). After 24 h of cold ischemia time (CIT), the gene expression levels of MMP-2 and TIMP-2 were found to be decreased in kidneys perfused with UW, whereas they were increased in kidneys perfused with EC. After warm ischemia, the levels of MMP-2 and TIMP-2 gene expression were increased in kidneys perfused with UW, whereas they were significantly lower in kidneys perfused with EC.

KIM-1, also known as hepatitis A virus cellular receptor 1 and T cell immunoglobulin- and mucin-domain-containing molecule, is considered to be a biomarker of ischemia-mediated renal injury during preservation (56). Its attractiveness stems from its specific expression pattern and function in renal proximal tubular cells. KIM-1 is a cell membrane glycoprotein (89), which is comprised of an extracellular portion, a six-cysteine immunoglobulin-like domain, two N-glycosylation sites and a thrombospondin-rich domain, which is characteristic of mucin-like O-glycosylated proteins (90). KIM-1 is predominantly localized on the apical surface of renal proximal tubular epithelial cells (91). It exhibits homology to a hepatitis A virus receptor (89). KIM-1 expression is typically absent under healthy conditions, but is activated under various conditions and diseases, such as cell dedifferentiation, ischemic injury (89), toxic injury (92), autosomal dominant polycystic kidney disease (93) and renal cell carcinoma (94). In the renal proximal epithelium, KIM-1 serves as a phosphatidylserine receptor that can recognize apoptotic cells and direct them into lysosomes (95). It can also bind to oxidized lipoproteins in injured cells. In addition, KIM-1 has been observed to mediate the transformation of epithelial cells into phagocytic cells (96).

A previous study on 671 kidneys retrieved from standard criteria donors (SCDs) and ECDs failed to demonstrate the utility of KIM-1 as a biomarker, predictor of DGF or long-term outcomes (56). KIM-1 concentrations were measured at the beginning and end of the perfusion. The mean difference between the KIM-1 concentrations at these points was not found to be associated with DGF. In an experimental model of NMP of discarded human kidneys, KIM-1 was found to be released at lower levels in kidneys with a higher urine output (64). In a previous study, in the DCDs kidney group, perfusate KIM-1 was proposed as an independent marker of DGF and was correlated with 3-month eGFR (57).

IL-18 is a cytokine that belongs to the IL-1 family and has been detected during hypothermic machine perfusion (25,56). IL-18 precursors are present in monocytes, macrophages, dendritic cells, intestinal cells, keratocytes, epithelial cells and renal tubular cells (97), in addition to being readily released by dying cells (97). Transcription of the IL-18 precursor is stimulated by the binding of Toll-like receptors to PAMPs, which then signals through the NF-κB pathway (97). Activation of the IL-18 precursor leads to its intracellular conversion into its mature form by caspase-1 (98). Extracellular activation mechanisms involve various proteases, such as neutrophil proteinase 3, granzyme B and chymase (99,100). In renal epithelial cells, the IL-18 precursor can be activated and converted into its mature form by the MMP meprin B (101).

IL-18 was originally identified to be an IFN-γ-inducing factor. It serves an important role in both naive and adaptive immunity (102). Among IL-12 and IL-15, which can upregulate IL-18 receptor expression, IL-18 induces IFN-γ production in CD4, CD8 T cells and natural killer (NK) cells (103). In addition, IL-18 can directly upregulate perforin- and FasL-dependent cytotoxicity in NK and CD8 T cells (103). IL-1β/IL-18 signaling initiates the transcription of several inflammatory factors, including GM-CSF, IL-4, histamine and TNFα, which facilitate leukocyte infiltration (104). IL-18, in the absence of IL-12 and IL-15, is responsible for the differentiation of naive T cells into T helper 2 cells, which produce IL-4 and IL-13, (105,106). Upon IL-18 stimulation, mast cells and basophils can produce IL-4 and IL-13 (107), whilst γδ T-cells produce IL-17 (97).

Previous studies on kidneys during storage using perfusate IL-18 could not confirm its predictive value for DGF (25,56). Its concentration was higher in samples from kidneys developing DGF, but no statistical significance could be found. In addition, although IL-18 associated with PNF in kidneys from DCDs, the diagnostic accuracy of IL-18 alone was poor (56). Similarly, in kidneys with a KDPI >80, IL-18 provided no predictive value for PNF, DGF or 1-year post-transplant outcomes according to another previous study (108). Table I shows a list of protein biomarkers detected using untargeted analysis methods. Fig. 2 shows molecules secreted during ischemia in HMP.