Future of non-invasive graft evaluation: A systematic review of proteomics in kidney transplantation

Abstract

BACKGROUND

Despite the developments in the field of kidney transplantation, the already existing diagnostic techniques for patient monitoring are considered insufficient. Protein biomarkers that can be derived from modern approaches of proteomic analysis of liquid biopsies (serum, urine) represent a promising innovation in the monitoring of kidney transplant recipients.

AIM

To investigate the diagnostic utility of protein biomarkers derived from proteomics approaches in renal allograft assessment.

METHODS

A systematic review was conducted in accordance with PRISMA guidelines, based on research results from the PubMed and Scopus databases. The primary focus was on evaluating the role of biomarkers in the non-invasive diagnosis of transplant-related complications. Eligibility criteria included protein biomarkers and urine and blood samples, while exclusion criteria were language other than English and the use of low resolution and sensitivity methods. The selected research articles, were categorized based on the biological sample, condition and methodology and the significantly and reproducibly differentiated proteins were manually selected and extracted. Functional and network analysis of the selected proteins was performed.

RESULTS

In 17 included studies, 58 proteins were studied, with the cytokine CXCL10 being the most investigated. Biological pathways related to immune response and fibrosis have shown to be enriched. Applications of biomarkers for the assessment of renal damage as well as the prediction of short-term and long-term function of the graft were reported. Overall, all studies have shown satisfactory diagnostic accuracy of proteins alone or in combination with conventional methods, as far as renal graft assessment is concerned.

CONCLUSION

Our review suggests that protein biomarkers, evaluated in specific biological fluids, can make a significant contribution to the timely, valid and non-invasive assessment of kidney graft.

Key Words: Proteomics; Kidney transplantation; Graft evaluation; Non-invasive diagnosis; Kidney graft function

Core Tip: In recent years, the role of diagnostic biomarkers in kidney transplantation has been emerged, particularly with proteomics being considered a powerful investigative tool. In this review, we critically analyze the spectrum of proteins exhibiting diagnostic accuracy in early, non-invasive evaluation of kidney grafts as well as the potent biological correlations between the variation of proteins detected in recipients experiencing some type of graft dysfunction or complication.

INTRODUCTION

Over the past decade, the number of solid organ transplantation operations performed worldwide has increased[1]. Kidney transplantation (KT) is the most common solid organ transplantation, with more than 25000 KT performed each year. Kidney grafts can be derived either from living donations or from donations after death[2]. Particularly in the latter category, grafts can derive from donors with brain or circulatory death (DCD), including extended criteria donors (ECD).

Despite the increasing number of KT performed annually, the long-term survival of the recipient and the graft have plateaued over the past decade[3]. The lack of a non-invasive sensitive diagnostic method routinely used in clinical practice limits the early non-invasive detection of possible graft damage. Moreover, kidney ECD and DCD allografts, typically derived from suboptimal donors, demonstrate an elevated propensity for graft loss, delayed graft function and kidney graft injury, thus requiring higher level of monitoring and evaluation[4]. Current clinical chemistry and biochemistry diagnostics in transplantation rely on a limited set of biomarkers alongside conventional methods. The last few years, the role of diagnostic biomarkers in KT has gained prominence with the inclusion of biomarkers evaluation in Banff classification for rejection, particularly in the diagnosis of antibody-mediated rejection, highlighting their potential importance in the future of KT[5]. Despite the recent breakthroughs, there is no single biomarker able to diagnose the complexity of transplant pathology, leaving a gap in transplantation diagnostics and highlighting the need for further exploration of the possible role of proteomics.

Proteomics are considered a powerful tool within the omics field, enabling the simultaneous comprehensive analysis of thousands of proteins expressed within cells, tissues, or organisms[6-8]. This approach allows for real-time evaluation of an individual’s state, distinguishing between health vs disease, and potentially predicting the susceptibility to specific complications. Proteomics analysis is a multi-analyte strategy, using cutting-edge technologies capable of assessing the complete set of proteins (proteome). While this approach is relatively new, the existing literature demonstrates promising results regarding sensitivity and specificity in the field of transplantation[8,9]. Proteomics in KT can be applied in kidney graft biopsies or in biological fluids, like serum and urine, providing a non-invasive multiplex, high throughput and amazingly sensitive diagnostic approach[10]. Furthermore, proteomics can also be applied to fluid from machine perfusion commonly used for the preservation of grafts, deriving from post mortem donation and particularly usually from DCD and ECD donors, facilitating early non-invasive diagnosis of transplant pathologies and preexisting graft damage[11]. Various proteomic approaches have been applied in clinical trials for diagnosing KT pathologies like rejection or BKV nephropathy, as well as for the detection of possible graft damage and the prediction of long- and short-term graft function[9]. Proteomics strategies provide a wide range of information, including identification of proteins in a sample at a given space, at a given moment, relative abundance levels of proteins or quantitative proteomics and analysis of post-translational modifications of those proteins mapping their specifically modified sites and their interactions. The increasing use of proteomics in everyday clinical practice is directly linked to recent technological advances in quantitation methods. Complex clinical samples may be evaluated more accurately, comprehensively, and promptly through recent technological breakthroughs, particularly in mass spectrometry (MS), sample preparation techniques, and computer data processing ability with bioinformatics and artificial intelligence or platforms, such as MaxQuant, Perseus, ProteoSign, UniReD, STRING and other[12-17].

The aim of our systematic review is to evaluate potential protein biomarkers, detected in biological fluids, for the evaluation of possible subclinical injury and fibrosis of the graft, as well as prediction of its short- and long- term function.

MATERIALS AND METHODS

The primary aim of this review was to identify the most recent advances in proteomic diagnostic applications regarding kidney graft evaluation by means of a systematic review, and to investigate possible correlations between the different proteins identified.

Eligibility criteria, information sources and search strategy

Following the PRISMA statement, we executed a systematic review of the literature for relevant studies using PubMed and Scopus databases by two independent reviewers. The search was conducted in all fields using the following keywords “kidney transplantation” or “renal transplantation” and “proteomics” or “protein biomarkers” until March 2024. The PubMed and Scopus databases were last accessed on 14 April 2024. To be eligible for screening, the studies had to meet the following criteria: (1) Identification of KT proteomic biomarkers in graft injury and function evaluation as the main objective; (2) Research performed in human urine or blood samples; (3) Application of specific methodologies for the quantitative and qualitative analysis of the samples; and (4) English language journal publication.

Study selection

We performed a systematic review of scientific publications that analyzed the protein profiles of kidney transplant samples or that detected specific proteins. All relevant studies were retrieved from the PubMed and Scopus databases without any restrictions on the date of publication. Studies were evaluated for inclusion by following a 2-phase selection process, as illustrated in Figure 1. All obtained studies were independently screened by title and abstract, by two independent reviewers. Articles that passed by mutual agreement were subsequently evaluated completely by the same authors to assess eligibility for inclusion, based on the inclusion and exclusion criteria. A study was included if its main purpose was to investigate the possible role of proteomics in the evaluation of kidney graft and function. We included cohort studies, clinical trials and original articles. Abstracts and studies with an unsuitable study design, such as narrative reviews, systematic reviews, meta-analysis, letters to the editor, book chapters, articles not published in English and articles where the full-text was not available, were excluded.

Figure 1 Prisma flow chart.

Data extraction and protein selection

Study characteristics, laboratory, and diagnostic specificity and sensitivity were extracted from the selected studies. Extracted study characteristics consisted of author, year of publication, study design, sample size, type of proteomic, population description and specific diagnosis or situation evaluated. Regarding proteins extracted from the articles, a table was created with annotations based on UniProt Database[18].

Categorization of biomarkers based on gene ontology annotations

We aimed to provide a comprehensive overview of all the proteomic biomarkers associated with graft evaluation and prediction of function in KT. Firstly, biomarkers were classified into two categories based on the biological sample from which they were measured: “blood-based”, “urine-based”, with the “blood-based” proteomics subdivided into “serum-based” and “plasma-based”.

The top-ranked proteins in the STRING database were used for gene set enrichment and annotation analyses (GSEA) based on gene ontology (GO), which is divided in the biological processes (BP), molecular functions (MF), and cellular components (CC), as well as functional and signaling pathways[14]. STRING facilitates GSEA by providing an enrichment analysis tool as part of its interface, using data from GO, KEGG, and other databases. The top ten categories on BP analysis were selected based on a combination of factors, including the pathophysiology mechanism correlated with KT, the protein count, and the false discovery rate. The top ten categories on MF and CC analysis where chosen based on protein count. Signaling pathways related to kidney dysfunction and kidney injury-associated key genes were retrieved from two different, freely available databases, REACTOME and WikiPathways[17-19]. Moreover, correlations based on disease-gene association and tissue expression were also used. The statistically significant GO terms and the significant pathways for hub-proteins were defined by the adjusted P values < 0.05.

RESULTS

Study selection and characteristics of included studies

The search strategy followed for selecting the eligible studies included in our systematic review is summarized in the flow chart shown in Figure 1, following PRISMA 2020. A total of 1754 potentially relevant manuscripts were identified through search of the PubMed and Scopus database searches. Of 159 manuscripts were removed before the abstract screening. Following the title and abstract screen, 87 manuscripts met the eligibility criteria and were subsequently assessed for inclusion. Of these, 23 didn't use appropriate quantitation techniques and 47 didn’t investigate the role of proteomics in the desired complications. In total, 17 studies met all the eligibility criteria and were included in this systematic review.

All manuscripts included in this review were published from 2013 onward, with most of them being published after 2018. The main characteristics of the eligible studies are presented in Table 1.

Table 1 Characteristics of the included studies.

Ref.	Year of publication	Biomarker name	Type of fluid	Method of quantification	No. of samples	Condition evaluated	Data availability
Buscher et al[35]	2023	HGF, IGFB, P6, CD40, CX3CL1, SPARCL1	Plasma	Mass cytometry, PEA	39	Allograft function and rejection status	Yes
Başak Oktay et al[20]	2022	CXCL10	Urine	LC-MS/MS, ELISA	18	Chronic allograft dysfunction	No
Jeon et al[26]	2022	RBP4	Urine	LC-MS/MS, ELISA	49	Evaluation of renal graft function	On request
Carreras-Planella et al[24]	2021	Vitronectin	Urine	LC-MS/MS, ELISA	59	Assessment of fibrosis in kidney graft	No
Navarrete et al[22]	2020	Proteinase 3	Urine	Serine hydrolase ABPP with Nano-RP LC–MS/MS	32	Subclinical rejection
Braun et al[36]	2020	PCK2	Urine	(Western blot, SDS-PAGE, IHC validation), nano-LC-MS/MS - both untargeted and targeted using PRM	22	Prediction of long-term graft outcome	Yes
Thorne et al[33]	2020	S-troponin and S-myoglobin	Plasma	LC-MS/MS, ELISA	225	Subclinical effects of remote ischemic reconditioning	On request
Bank et al[31]	2019	TIMP2	Urine	sandwich ELISA	83	Prediction of the occurrence and duration of functionally defined delayed graft function	Yes
Al-Nedawi et al[25]	2019	TFIID subunit 1, 3-hydroxy-3-methylglutaryl coenzyme A reductase, FN1 protein, Spectrin beta chain, Protein AMBP, Alpha-1-acid glycoprotein 2, Dynein heavy chain 3, axonemal, Insulin receptor substrate 1, Prothrombin, Alpha-1-acid glycoprotein 1, Melanotransferrin, Catalase, DIS3-like exonuclease 1, Kinetochore-associated protein 1, Keratin, type II cuticular Hb6, Neuroblast differentiation-associated protein AHNAK, Apolipoprotein E, Angiotensinogen, Vitamin D-binding protein, EGF-containing fibulin-like extracellular matrix protein 1, Keratin, type II cytoskeletal 3, DNA repair protein RAD50, Receptor-type tyrosine-protein phosphatase, Apolipoprotein A-II, Transthyretin, Hemoglobin subunit beta	Plasma	(SDS-PAGE), MS/MS	30	eGFR correlation and graft function evaluation	No
Mockler et al[29]	2018	CXCL10	Urine	LC-MS/MS (to measure creatinine), sandwich ELISA (to measure CXCL10 and CCL2)	59	Correlation with eGFR and long term renal graft outcome	No
Williams et al[32]	2017	C4BPA, alpha, guanylin, immunoglob- ulin superfamily member 8, and serum amyloid P- component.	Urine	TUPA (scheduled LC-MRM this is a type of MS), sandwich ELISA	52	Diagnosis of Delayed graft function	Yes
Hirt-Minkowski et al[27]	2016	CCL2, CXCL10	Urine	ELISA	185	Prediction of long-term renal graft outcome	Yes
Hirt-Minkowski et al[28]	2015	CXCL10	Urine	sandwich ELISA	154	Long-term renal allograft outcome evaluation	Yes
Ho et al[23]	2016	MMP7	Urine	LC-MS/MS, ELISA	17	Subclinical graft injury and inflammation	Yes
Sigdel et al[21]	2013	ACAN, CLEC14A, GLB1, LGALS9B, LMAN2, MRC2 SEL, CD27, DPEP1, and F2	Urine	iTRAQ labeling, LC-MS/MS, ELISA	142	Evaluation of chronic allograft injury	Yes
Welberry Smith et al[30]	2013	Aminoacylase-1	Serum	LC-MS/MS, sandwich ELISA	194	Prediction of long-term outcome in patients with delayed graft function	Yes
Schmidt et al[34]	2013	Chitinase-3-like protein 1	Urine	MS	78	Evaluation of delayed graft injury and prediction of long-term renal graft outcome	Yes

LC-MS/MS: Liquid chromatography - tandem mass spectrometry; MRM: Targeted; ELISA: Enzyme-linked immunosorbent assay; HGF: Hepatocyte growth factor; CX3CL1: Fractalkine; IGFBP6: Insulin growth factor binding protein 6; CXCL10: C-X-C motif chemokine 10; SPARCL1: SPARC-like protein 1; RBP4: Retinol-binding protein 4; PCK2: Phosphoenolpyruvate carboxykinase 2; TIMP2: Tissue inhibitor of metalloproteinases-2; TFIID: Transcription initiation factor; C4BPA: C4b-binding protein alpha chain; CCL2: Chemokine ligand 2 ; MMP7: Matrix Metalloproteinase-7.

Sample and population characteristics

The median number of samples was 59 (interquartile range 117), whereas 12 (70.58%) studies had under 100 samples. Out of 17 proteomics studies, 13 used samples derived from urine and four used blood-based samples. Three of the blood-based samples were plasma and one study employed serum samples. Figure 2 shows quantitative data on biological sample selection in all included samples.

Figure 2 Quantitative data about the biological sample selection in all the included samples.

Evaluation techniques

The most commonly used methodology in the 16 articles reviewed was liquid chromatography - tandem MS (LC-MS/MS), targeted multiple reaction monitoring (MRM) or untargeted (discovery). This combines two techniques consecutively: Liquid chromatography and tandem MS, to create a powerful identification and quantitation tool for biomolecules with both high sensitivity and specificity. However, it is essential to consider the potential differences in detection sensitivity between the targeted MRM approach and the unbiased, system-wide, untargeted discovery approach in LC-MS/MS. MRM is optimized for specific analytes, which can result in higher sensitivity and precision, allowing for more accurate quantitation of (especially) low-abundance proteins or peptides. On the other hand, untargeted LC-MS/MS methods aim to detect a wide range of analytes in a single run, which can sometimes be at the expense of sensitivity for singular analytes. These differences can potentially lead to inconsistencies when comparing results between targeted and untargeted approaches. A few papers used both MRM and discovery LC-MS/MS, proving an effective way to mitigate and complement such sensitivity differences and validate results.

This method was most commonly paired with a direct or sandwich enzyme-linked immunosorbent assay (ELISA) for validation. ELISA uses a single antibody directly linked to an enzyme to detect and quantify a target protein in a given sample, while sandwich ELISA uses both a capture and a detection antibody for added sensitivity and specificity. Other methods used include targeted urine proteome assay, which is a type of multiple reaction monitoring (MRM), isobaric tags for relative and absolute quantitation labeling, as well as serine hydrolase activity-based protein profiling and proximity extension assay.

Biological samples proteome

In the total of 17 studies, 58 proteomic biomarkers were evaluated (Table 2). Of these, one protein (CXCL10) overlapped between different studies, as seen on the heatmap in Figure 3. Among 58 proteins identified, 23 were detected in urine samples, 33 were detected in plasma samples and two were detected in serum samples, with CXCL10 being detected both in urine and serum samples. Out of 58 proteomic biomarkers, 39 were found to be elevated compared to control/healthy samples, 12 were found to be down-regulated, six were related to better prognosis, one was detected only in samples from injured kidney grafts, and one appeared with increased activity in samples derived from complicated KT.

Figure 3 Heatmap representation of proteins detected across various studies related to kidney transplantation. Each column represents a different study or dataset, and each row represents a specific protein. Green color indicates the protein detection within each study.

Table 2 The table contains the protein's name, UniProt number, cellular location, and code gene.

Name	UniProt code	Cellular location	Sample type	Status	Code gene
Hepatocyte growth factor	P14210	Membrane, extracellular space	Plasma	Elevated	HGF
Insulin growth factor binding protein 6	P24592	Secreted	Plasma	Elevated	IGFBP6
CD40	P25942	Cell membrane and secreted	Plasma	Elevated	CD40
Fractalkine	P78423	Cell membrane	Plasma	Elevated	CX3CL1
SPARC-like protein 1	Q14515	Secreted, extracellular space and extracellular matrix	Plasma	Elevated	SPARCL1
C-X-C motif chemokine 10	P02778	Secreted, plasma membrane	Urine	Elevated	CXCL10
Retinol-binding protein 4	P02766	Cell wall, organelle, Secreted	Urine	Elevated	RBP4
Phosphoenolpyruvate carboxykinase 2	Q16822	Mitochondrion, ribosome	Urine	Elevated	PCK2
S troponin	P19429	Cytosol and sarcomere	Serum	Elevated	TNNI3
S myoglobin	P02144	Cytoplams and sarcoplasm	Serum	Elevated	MB
TIMP2	P16035	cytoplasmic vesicle, organelle, Secreted	Urine	Elevated	TIMP2
CCL2	P13500	Secreted	Urine	Elevated	CCL2
MMP7	P09237	Secreted, extracellular space, extracellular matrix	Urine	Elevated	MMP7
Aminoacylase-1	Q03154	Cytoplasm	Serum	Elevated	ACY1
Chitinase-3-like protein 1	P36222	Cytoplasm, endoplasmic reticulum, secreted	Urine	Elevated	CHI3 L1
Protein AMBP	P02760	Extracellular region, secreted	Plasma	Elevated	AMBP
Alpha-1-acid glycoprotein 2	P19652	Extracellular region, secreted	Plasma	Elevated	ORM2
Dynein heavy chain 3	Q8TD57	Cytoskeleton, cilium axoneme	Plasma	Elevated	DNAH3
Insulin receptor substrate 1	P06213	Cytosol, nucleus, plasma membrane, caveola	Plasma	Elevated	INSR
Catalase	P04040	Peroxisome	Plasma	Elevated	CAT
DIS3-like exonuclease 1	Q8TF46	Cytoplasm	Plasma	Elevated	DIS3 L
Kinetochore-associated protein 1	P50748	Cytoskeleton, spindle, nucleus, cytoplasm	Plasma	Elevated	KNTC1
Apolipoprotein E	P02649	Extracellular region, secreted	Plasma	Elevated	APOE
Angiotensinogen	P01019	Extracellular region, secreted	Plasma	Elevated	AGT
Vitamin D-binding protein	P02774	Extracellular region, secreted	Plasma	Elevated	GC
EGF-containing fibulin-like extracellular matrix protein 1	Q580Q6	Extracellular region, secreted	Plasma	Elevated	EFEMP1
C4b-binding protein alpha chain	P04003	Secreted	Urine	Elevated	C4BPA
Serum amyloid P-component	P02743	Secreted	Urine	Elevated	APCS
Aggrecan core protein	P16112	Secreted, extracellular space, extracellular matrix	Urine	Elevated	ACAN
C-type lectin domain family 14 member A	Q86T13	Membrane	Urine	Elevated	CLEC14A
Beta-galactosidase	P54803	Cytoplasm, lysosome	Urine	Elevated	GLB1
Galectin-9B	Q3B8N2	Cytosol, nucleus	Urine	Elevated	LGALS9B
Vesicular integral-membrane protein VIP36	Q12907	Endoplasmic reticulum, golgi apparatus, membrane	Urine	Elevated	LMAN2
C-type mannose receptor 2	Q9UBG0	Membrane	Urine	Elevated	MRC2
Protein sel-1 homolog 1	Q9UBV2	Endoplasmic reticulum membrane	Urine	Elevated	SEL1 L
CD27 antigen	P26842	Membrane	Urine	Elevated	CD27
Dipeptidase 1	Q9H4A9	Cell projection, microvillus membrane, cell membrane	Urine	Elevated	DPEP1
Transcription factor E2F2	Q14209	Nucleus	Urine	Elevated	E2F2
Keratin, type II cytoskeletal 3	P12035	Intermediate filament, keratin filament cytosol, exosome	Plasma	Decreased immediately, elevated 1 month after	KRT3
DNA repair protein RAD50	Q92878	Nucleus	Plasma	Decreased immediately, elevated 1 month after	RAD50
Receptor-type tyrosine-protein phosphatase	P23469	Integral component of membrane	Plasma	Decreased immediately po, elevated 1 month after	PTPRE
Vitronectin	P04004	Endoplasmic reticulum, golgi apparatus, organelles, secreted	Urine	Detected only in fibrotic grafts	VTN
proteinase 3	P24158	Lysosome, cytosol, plasma membrane, cytoplasmic vesicle, organelle, secreted	Urine	Increased activity	PRTN3
Transcription initiation factor TFIID subunit 1	P21675	Nucleus	Plasma	Decreased	TAF1
3-hydroxy-3-methylglutaryl coenzyme A reductase	P04035	Endoplasmic reticulum, ER membrane	Plasma	Decreased	HMGCR
Fibronectin	P02751	Extracellular region, secreted	Plasma	Decreased	FN1
Spectrin beta chain, non-erythrocytic 1	Q01082	Cytoskeleton	Plasma	Decreased	SPTBN1
Apolipoprotein A-II	V9GYM3	Extracellular region, secreted	Plasma	Decreased	APOA2
Transthyretin	E9KL36	Extracellular region, secreted	Plasma	Decreased	HEL111
Hemoglobin subunit beta	P68871	Cytosol, extracellular or secreted: Blood microparticle, exosome	Plasma	Decreased	HBB
Guanylin	Q02747	Secreted	Urine	Decreased	GUCA2A
Immunoglobulin superfamily member 8	Q969P0	Cell membrane	Urine	Decreased	IGSF8
Prothrombin	P00734	Extracellular region, secreted	Plasma	Elevated in good prognosis recipients	F2
Alpha-1-acid glycoprotein 1	P02763	Secreted	Plasma	Elevated in good prognosis recipients	ORM1
Melanotransferrin Isoform 1	P08582	Plasma membrane, GPI-anchor	Plasma	Elevated in good prognosis recipients	MELTF
Apolipoprotein A-I	P02647	Extracellular region, secreted	Plasma	Elevated in good prognosis recipients	APOA1
Keratin, type II cuticular Hb6	O43790	Cytoskeleton, keratin filament, cytosol extracellular region or secreted, exosome	Plasma	Elevated in good prognosis recipients	KRT86
Neuroblast differentiation-associated protein AHNAK	Q09666	Cytoskeleton, centrosome	Plasma	Elevated in good prognosis recipients	AHNAK

HGF: Hepatocyte growth factor; CX3CL1: fractalkine; IGFBP6: Insulin growth factor binding protein 6; CXCL10: C-X-C motif chemokine 10; SPARCL1: SPARC-like protein 1; RBP4: Retinol-binding protein 4; PCK2: Phosphoenolpyruvate carboxykinase 2; TIMP2: Tissue inhibitor of metalloproteinases-2; C4BPA: C4b-binding protein alpha chain; CCL2: Chemokine ligand 2; MMP7: Matrix Metalloproteinase-7.

Protein categorization and enrichment analysis based on GO terms

All proteins deriving from healthy and pathological samples were subjected to GO enrichment analysis, including BP, CC and MF terms. A separate analysis of proteins found upregulated in samples from healthy and complicated KT was also performed. A total of 58 proteins were identified using their UniProt accession number. Figure 4 summarizes the top ten significant and enriched GO terms of BP, CC, and MF based on increased proteins in samples from complicated KT recipient cases, with the false discovery rate, a measure that describes how significant the enrichment is by showing p-values corrected for multiple testing within each category, using the Benjamini–Hochberg procedure, mentioned in the parenthesis. In the BP analysis performed based on increased proteins found in samples of KT recipients with some type of graft injury or complication, the majority of related proteins were involved in catalytic activity, immune response and stress defense. In terms of CC analysis, most proteins were located in the extracellular space, extracellular exosome and extracellular matrix, while a significant percentage of proteins were secreted from the cells. Of the MF terms, signaling receptor binding and molecular function regulator activity were the most important functions. In the analysis of elevated protein abundance levels correlated with good outcome, due to the limited protein number, only CC analysis was performed, and the results showed that all six proteins were located in the extracellular space, with three of them being particularly blood microparticles.

Figure 4 Gene ontology analysis of upregulated proteins detected on samples of kidney transplant recipients with graft injury or dysfunction. False discovery rate is mentioned in the parenthesis.

Application of proteomics in kidney graft evaluation

The previously discussed proteomics findings exhibit both diagnostic and predictive value. Their diagnostic value relies on the correlation between the increase or decrease of specific proteins detected in biological fluids of KT recipients and the histological alterations observed in conventional biopsies, often accompanied by other functional biomarkers, such as creatinine. Among the 17 studies included in this systematic review, five studies investigated the potential connection between protein biomarker non-invasive assessment and conventional kidney biopsy results[20-36]. Out of those five studies, Başak Oktay et al[20] and Sigdel et al[21] focused on diagnosing chronic allograft dysfunction through proteomic assessment, while Navarrete et al[22], and Ho et al[23] investigated subclinical rejection and injury, and Carreras-Planella et al[24] explored possible correlation between specific protein biomarkers with fibrosis levels observed in kidney graft biopsies.

Evaluation and prediction of function of the kidney graft can be achieved by the evaluation of protein biomarkers in correlation with conventional methods, such as estimated glomerular filtration rate (eGFR) and creatinine levels assessment. Specifically, 12 studies correlated proteomic biomarkers with conventional laboratory kidney function evaluation techniques, in order to assess or predict its function[25-36]. Al-Nedawi et al[25] and Jeon et al[26] investigated potential links between bad outcome kidney transplant cases, based on the eGFR of the recipients, and proteomic assessment. For predicting long-term graft function, Hirt-Minkowski et al[27,28], Mockler et al[29], and Braun et al[36] emphasized the importance of measuring proteins at specific time points. Similarly, Welberry Smith et al[30] highlighted that measurement of specific protein biomarkers on days 1 and 2 post transplantation can be predictive of this complication. Bank et al[31] also correlated specific proteomic panels with eGFR and function of the kidney graft, particularly focusing their study on short-term graft function evaluation on days 1 and 10 after transplantation. Furthermore, Williams et al[32] also evaluated the same complication by correlating eGFR with specific biomarkers assessed 12-18 hours after transplantation. An additional important investigation was conducted by Thorne et al[33] where ischemic reconditioning, a situation that is part of KT process, was studied from the proteomics perspective. Moreover, long-term graft function prediction can also have significant clinical importance according to a study performed by Schmidt et al[34]. Lastly, Buscher et al[35] correlated specific plasma proteomic panels with allograft function based on creatinine levels, underscoring the diagnostic superiority of a proteomic panel in contrast to specific proteins in graft function evaluation.

DISCUSSION

Renal transplantation is the most common transplant operation and the only effective treatment for end-stage kidney disease, offering recipients a significantly improved quality of life compared to dialysis. Despite the first successful kidney transplant operation being performed more than 70 years ago, graft-related complications still exist due to various causes. Current clinical practice for monitoring KT recipients relies on conventional markers such as serum creatinine, urine volume, proteinuria, and other laboratory tests. Although these biomarkers can indicate some form of kidney injury, they typically signal allograft damage only after it has occurred. Biopsies remain the primary tool for diagnosis, offering high sensitivity and specificity, but they are invasive procedures associated with major complications and potential graft loss[37]. Proteomic approaches offer exciting groundbreaking discoveries connecting basic translational research with transplant everyday clinical practice, enabling early assessment and prediction of various possible complications in a non- or minimal invasive way.

Proteomic data analysis, consisting of identification, relative quantitation and functional enrichment and signaling pathways analysis, represents a powerful strategy for deciphering intricate molecular activities and functional pathways associated with a particular group of proteins. Notably, in this systematic review among the GO terms, those correlated with immunological activity were significantly enriched, as seen in Figure 4. Specifically, the terms of regulation of immune system process, inflammatory response, regulation of cytokine production, response to IL-1, neutrophil migration and lymphocyte chemotaxis were enriched in total by 16 proteins in the GO BP analysis of upregulated proteins found in samples of KT recipients with complications, as shown on the bitmap on Figure 5. Analysis based on Reactome and WikiPathways databases has also underscored the association with immune response, notably IL-1. Given the strong relationship between KT and the immune system, both short- and long-term complications are often attributed to abnormal immune responses[38,39]. Despite the latest major advancements in immunosuppression, even subclinical graft injury can jeopardize the function of the graft[40,41]. Particularly, the significantly enriched IL-1 highlights its importance as a cytokine target for potential immunosuppressive therapies[42-44]. The development of IL-1 target agents began in 1993 with the introduction of anakinra, a recombinant form of the naturally occurring IL-1 receptor antagonist, with major clinical applications due to its excellent therapeutics being related to its safety, short half-life and multiple routes of administration[45].

Figure 5 Protein interaction network analysis. A: A protein interaction network of upregulated proteins detected in biological fluids of recipients with complicated transplantations. The proteins related to immune biological pathways are highlighted; B: A protein interaction network of proteins related to normal kidney function.

Another notable finding derived from the Reactome and WikiPathways analyses of proteins originating from pathological samples is the association with elastic fibers and wound healing pathways. These pathways are correlated with graft fibrosis, a form of subclinical graft injury resulting in long term dysfunction of the graft[46-49]. This correlation is also supported by the diseases-gene association, as well as highlighted by Carreras-Planella et al[24] in their study[24]. Specifically, Metalloproteinase-9 and Metalloproteinase inhibitor 2, two proteins found to be enriched in pathways related to fibrosis, have been reported to be correlated with kidney graft injury, and they have been also proved to have predictive value for the long-term graft function[50,51]. Moreover, Fibronectin, which is another protein omic related to fibrosis, has been correlated with a rare inherited kidney disorder, named Fibronectin glomerulopathy, which results in impaired stress fiber formation, even after successful transplantation[52].

One intriguing correlation in the disease-gene association was one associating four proteins of our analysis with amyloidosis. Amyloidosis is a rare condition caused by abnormal folding of proteins, which results in complications related to many organs, including the kidneys[53]. Regarding KT, amyloidosis can be caused, either as a result of recurrence of the preexisting condition, or as a newly developed complication leading to graft loss[54,55].

Another significant association found was with blood coagulation proteins. In particular, 10 upregulated proteins in complicated cases of KT demonstrated commonality with cardiovascular diseases, among these being Apolipoprotein E. Apolipoprotein E is a multifunctional protein central to lipid metabolism, contributing to atherosclerosis and consequent damage of the arteries[56]. The role of Apolipoprotein E in chronic kidney disease has been well documented, where lipid and non-lipid pathways are implicated in the disease onset and progression[57]. Subclinical graft injury of the graft and subsequent dysfunction are directly linked with damage of the vessels of the kidney graft. Particularly, specific gene polymorphisms, such as E4 allele have been related to poor long-term transplant outcome[58,59].

End-stage kidney failure is recognized as the primary reason for KT[60]. Myoglobin, Activation peptide fragment 1 and Angiotensin are proteins associated with kidney failure, and they were found to be enriched in the disease-gene association STRING analysis. Angiotensin is a well-studied biomarker related to kidney graft injury with the mechanism of the intrarenal RAS activation via inflammation and oxidative stress[61]. Myoglobin is linked to rhabdomyolysis, a pathological situation, resulting to serious graft damage[62].

In addition to the increased protein abundance levels found in complicated cases of KT, it is noteworthy to emphasize the possible protective or predictive role of proteins found to be elevated abundance levels in good prognosis KT recipients. Notably, Prothrombin, Alpha-1-acid glycoprotein 1, Melanotransferrin Isoform 1, Apolipoprotein A-I and Keratin type II cuticular Hb6 were reported to be increased in good prognosis patients by Al-Nedawi et al[25]. Prothrombin, Alpha-1-acid glycoprotein 1 and Apolipoprotein A-I also all interact, as indicated in Figure 5, due to their co-expression in events related to blood cascade. In the context of KT and kidney disease, Alpha-1-Acid Glycoprotein has protective role to kidney injury, while apolipoprotein A has been related to lower prevalence of CKD, thus related to good kidney function[63,64].

Out of the 58 protein biomarkers evaluated in the total of 17 studies, CXCL10 is the only one mentioned in more than one study. Particularly, the role of this protein biomarker in KT evaluation is studied by Başak Oktay et al[20], Mockler et al[29] and Hirt-Minkowski et al[27,28]. CXCL10 has been reported to be elevated in pathological urine or serum samples of KT recipients. This protein is a cytokine that belongs to the CXC chemokine family and it is secreted by activated B cells, monocytes/macrophages, endothelial cells, and fibroblasts in response to interferon gamma (IFN-γ). CXCL10 plays a pivotal role in inflammation, due to its chemoattractive effects for monocytes/macrophages, T cells, natural killer (NK) cells, and dendritic cells via cell surface chemokine receptor CXCR3[65]. Involvement of CXCL10 in renal diseases is well-documented, with the biomarker having a pivotal role in Mesangial Proliferative Glomerulonephritis, acute kidney injury and lupus nephritis[66]. Its significant role in renal disease pathophysiology is supported by the observation that renal resident mesangial cells, renal tubular epithelial cells, podocytes, endothelial cells, and infiltrating inflammatory cells express CXCL10 under inflammatory conditions[66].

Lastly, it is important to highlight the choice of quantitation methods in different proteomics studies. As mentioned in the results section, the most used methodology is LC-MS/MS, targeted on untargeted. Using an untargeted MS-based approach with a targeted follow-up, as was done in LC-MS/MS followed by ELISA, has an added benefit of eliminating bias in protein selection, further validating the results, while ensuring that increased quantitation of specific proteins was accurately measured and recorded[67]. On the other hand, using a targeted approach allows research bias to play a part in the experiment, so only previously recognized proteins with potential for biomarker status can be used. Using non-MS based methods guarantees this targeted approach, however it does increase specificity, as well as ease the load for data analysis and management.

Our study has some specific limitations. The primary limitation was the heterogeneity among enrolled studies, with big heterogeneity of sample, population, time of sample collection and researched pathology being reported, which limited the possible statistical analysis possibility. Moreover, the study used a comprehensive seed proteomic dataset from a systematic literature review, including databases of PubMed and Scopus. Additionally, during the proteomic data collection, only association with graft injury and function were considered while information regarding the source of the graft (living or cadaveric donation, brain or circulatory death donation) were not taken into consideration in the separate analysis. Lastly, the variation recorded in studies in detection sensitivity between LC-MS/MS and MRM methods used can pose substantial challenges. LC-MS/MS provides wide detection but has variable quantitation, while MRM delivers very sensitive and accurate tailored quantitation. These differences may cause discrepancies in comparative analyses, adding bias, hindering data standardization and integration. Advanced statistical models and thorough standardization are required to harmonize data and produce accurate, dependable outcomes across research.

CONCLUSION

This systematic review showed that very few proteomic-level studies have been conducted to identify the key proteins associated with KT complications, and particularly with subclinical injury and prediction of graft function. Nevertheless, to the best of our knowledge, we were unable to locate any gene expression data regarding KT and non-invasive graft evaluation. We highlight the importance of new research studies to be held in order to identify the best possible panel of proteins for the early non/minimal invasive detection of different complications, as well as the best possible method of evaluation.