Differentiating the Whole Urine and Urine Supernatant Protein Profiles of Preterm Infants via Urine Proteomics

With the continuous development of neonatal intensive care technology, the survival rate of newborns, especially preterm infants and low birth weight infants, has been increasing [1]. According to the global preterm birth data released by WHO in 2018, the global preterm birth rate is about 10.6%, with China ranking second at about 7.1% [2, 3]. It is accompanied by the development and improvement of the structure and function of various systems of the body after birth. However, the organs of the body begin to play their functions in advance due to preterm birth, which is bound to cause damage to its long-term structure and function. Epidemiological studies indicate that preterm birth is a significant risk factor for long-term systemic diseases [4‒7]. However, the invasiveness of routine clinical and laboratory tests and the retardation in diagnosis of the disease are serious problems in the intensive care of preterm infants.

In terms of the renal system, preterm birth and low birth weight are associated with increasing risk of long-term chronic kidney disease such as glomerulonephritis, glomerulosclerosis, tubular damage, and renal vasculopathy [8, 9]. Crump et al. [10] found that the incidence of chronic kidney diseases in preterm infants was twofold higher than that of full-term infants. However, the most progression from kidney injury to end-stage kidney disease is insidious and progressive [11]. It has no specific symptoms, and early diagnosis and treatment are currently difficult in clinical practice. Therefore, early diagnosis and early detection of renal pathological changes in children with high-risk factors such as preterm birth have important clinical value. It is of great significance for evaluating disease progression, early intervention and treatment, improving prognosis, and quality of life of children. Currently, routine kidney biopsy is helpful in identifying the pathological changes of the kidney in children and assessing the progression of kidney disease [12‒14]. However, renal biopsy is a traumatic examination, and the compliance of children with repeated renal biopsy follow-up changes is poor. So, it will be of great value to seek some noninvasive biological indicators that can predict the pathological state of kidney.

With the completion of the Human Genome Project, life science research has entered the post-genome era of genomics, proteomics, metabolomics, etc. [15]. This allows genes and gene products to act as biomarkers. Compared with the unique and relatively stable genes in the genome, the proteome is cell- and tissue-specific and constantly changes over time, directly reflecting different life states of the body. Proteomics plays an important role in screening biomarkers for diagnosis and prognosis of diseases in clinical studies. In neonatal care, preterm infants account for a large proportion of infants with small gestational weeks and low birth weight. Noninvasive methods to obtain biological samples for diagnosis and longitudinal continuous detection to evaluate treatment effect and prognosis have become a good choice in clinical practice. At present, the noninvasive biological samples of children used for protein profiling mainly include urine, oral scrapes, and exhaled condensation [16‒18]. For newborns, a large amount of urine can be obtained without invasion, and urine can not only reflect the pathological changes of the urinary tract and kidneys but also reflect the pathophysiological changes of other organs of the body. Urine proteins has been shown to remain stable long enough for reliable proteomic analysis. Among them, proteomic detection of infant urine samples is widely used to predict kidney injury, necrotizing enterocolitis (NEC), bronchopulmonary dysplasia, and other diseases, as well as to monitor the maturity and development of urinary system [19‒21].

For urine proteomics, urine sample is the basis of research, and whether the treatment of urine sample before detection has an impact on the screening of biomarkers is a very important and meaningful subject. Currently, there is no unified standard for the preparation of urine samples in the study of neonatal urine proteomics. Generally speaking, there are two main treatment methods, one is whole urine specimen [19, 22] and the other is urine supernatant specimen after centrifugation [23]. However, no studies have clarified the circumstances under which whole urine samples should be used for analysis and the circumstances under which urine supernatant after centrifugation should be used for urine proteomic analysis, and the differences between these two urine sample processing methods in urine protein profiles are not yet clear. The urine sample contains 97% water, and the rest includes organic components such as cellular components, tubular components, proteins, fats, microorganisms, and inorganic components such as sodium, potassium, calcium, magnesium, sulfate, phosphate, and so on [24]. The composition of urine samples was different in different disease status. The amount and solubility of protein in urine vary with different disease states of the body. In the urine of healthy people, protein will not appear excessive aggregation and precipitation, but in some disease states, protein in urine will appear excessive aggregation and spontaneous precipitation and other uncontrollable phenomena [25]. Theoretically, the whole urine specimen contains more comprehensive components, which can better reflect the physiological and pathological status of the body. However, the whole urine sample has some disadvantages, such as protein degeneration, cell disruption, and bacterial multiplication [26, 27]. Urine supernatant was obtained by centrifugation, and it removes the effect of visible components such as cells and tubes on proteins. However, there is also the disadvantage of loss of large molecular weight proteins, especially diagnostic related proteins, and potential protein biomarkers in the process of urine specimen centrifugation. Therefore, based on the differences in the treatment methods of urine samples in the urine protein profile of preterm infants, urine samples of preterm infants were divided into whole urine samples and urine supernatant sample, and exploratory analysis was conducted on urine samples of preterm infants treated with the two methods to clarify the differences in the protein profiles of urine samples prepared by the two methods.

This study was approved by the Tianjin Central Hospital of Gynecology and Obstetrics Institutional Review Board (No. 2020KY049). All study protocols followed the Declaration of Helsinki and obtained the written informed consent of all participants’ parents prior to inclusion in this study.

Five preterm infants, all of whom were less than 28 weeks old, were contained in the study. The urine samples were collected from preterm infants at their corrected gestational age reached more than 37 weeks when kidneys have completed nephrogenesis. The urine samples of each preterm baby were divided into two parts, one was reserved for whole urine sample, and the other was processed to obtain urine supernatant sample.

A sterile collection bag was used to collect urine for each preterm infant. Each urine collection bag was monitored every 30 min. After collection, the urine samples were immediately transferred to two sterile centrifuge tubes, one of which was reserved for whole urine samples and stored at −80°C, while the other one was sent to the laboratory and centrifuged at 4,000 rpm for 15 min. The urine supernatant was taken and stored in a new sterile centrifuge tube at −80°C for further detection.

The protein was quantified with a BCA Protein Assay Kit (Bio-Rad, USA). Protein (300 μg for each sample) digestion was performed with FASP method described by Wisniewski, Zougman et al. [28]. Briefly, the detergent, DTT (100 mm), and IAA (100 µL, 50 mm IAA in UA), in UA buffer (200 µL, 8 m urea, 150 mm Tris-HCl, pH 8.0) was added to block reduced cysteine. Finally, the protein suspension was digested with trypsin (40 µL, 6 µg trypsin in 40 µL NH₄HCO₃ buffer) about 16–18 h at 37°C. The peptides were collected by centrifugation at 12,000 g for 10 min. The peptide was desalted with C18 Stage Tip for further LC-MS analysis. The concentrations of peptides were determined with OD280 by NanoDrop One device. We established a minimum threshold of one peptide for protein identification to ensure data reliability, given the enhanced accuracy and resolution of contemporary mass spectrometers.

Liquid chromatography-tandem mass spectrometry (LC-MS/MS) were performed on a Q Exactive Plus mass spectrometer coupled with Easy 1200 nLC (Thermo Fisher Scientific). Peptide was first loaded to a trap column (100 μm*20 mm, 5 μm, C18, Dr. Maisch GmbH, Ammerbuch, Germany) in buffer A (0.1% formic acid in water). Reverse-phase high-performance liquid chromatography separation was performed with the EASY-nLC system (Thermo Fisher Scientific, Bremen, Germany) using a self-packed column (75 μm × 150 mm; 3 μm ReproSil-Pur C18 beads, 120 Å, Dr. Maisch GmbH, Ammerbuch, Germany) at a flow rate of 300 nL/min. The reverse-phase high-performance liquid chromatography mobile phase A was 0.1% formic acid in water, and B was 0.1% formic acid in 95% acetonitrile. The liquid phase gradient was set as follows: 0 min–2 min, B liquid linear gradient from 5% to 8%; 2 min–90 min, B liquid linear gradient from 8% to 23%; The linear gradient of B solution was from 23% to 40% at 90 min–100 min. Linear gradient of B liquid from 40% to 100% in 100 min–108 min; The B solution was maintained at 100% from 108 min to 120 min. Peptides were separated and analyzed by DDA (data-dependent acquisition) mass spectrometry on a Q Exactive HF-X mass spectrometer (Thermo Scientific). Peptide were eluted over 120 min with a linear gradient of buffer B. MS data were acquired using a data-dependent top 20 method dynamically choosing the most abundant precursor ions from the survey scan (300–1,800 m/z) for HCD fragmentation. The full MS scans were acquired at a resolution of 60,000 at m/z 200 and 15,000 at m/z 200 for MS/MS scan. The maximum injection time was set to for 50 ms for MS and 50 ms for MS/MS. Normalized collision energy was 28, and the isolation window was set to 1.6 m/z.

The MS data were analyzed using MaxQuant software version 1.6.0.16. MS data were searched against the UniProtKB human database. The trypsin was selected as digestion enzyme. The maximal two missed cleavage sites and the mass tolerance of 4.5 ppm for precursor ions and 20 ppm for fragment ions were defined for database search. Carbamidomethylation of cysteines was defined as fixed modification, while acetylation of protein N-terminal, oxidation of methionine was set as variable modifications for database searching. The database search results were filtered and exported with <1% false discovery rate at peptide-spectrum-matched level and protein level, respectively Label-free quantification was carried out in MaxQuant using intensity determination and normalization algorithm as previously described [29]. The “LFQ intensity” of each protein in different samples was calculated as the best estimate, satisfying all of the pairwise peptide comparisons, and this LFQ intensity was almost on the same scale of the summed-up peptide intensities. The quantitative protein ratios were weighted and normalized by the median ratio in MaxQuant software. Only proteins with fold change ≥1.5-fold and a p value <0.05 were considered for significantly differential expressions.

Analyses of bioinformatics data were carried out with Perseus software (Version 1.5.5.3) [30], Microsoft Excel and R statistical computing software. Hierarchical clustering analysis was performed with the pheatmap package, which is based on the open-source statistical language R25, using Euclidean distance as the distance metric and complete method as the agglomeration method. To annotate the sequences, information was extracted from UniProtKB/Swiss-Prot, Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Ontology (GO). GO and KEGG enrichment analyses were carried out with the Fisher’s exact test. Enriched GO and KEGG pathways were nominally statistically significant at the p < 0.05 level.

Urine samples were collected from 5 preterm infants at the Neonatology Department of Tianjin Central Hospital of Gynecology and Obstetrics. The clinical and demographic characteristics of preterm infants are listed in Table 1. The demographic characteristics including gestation age at birth, birth weight, sex, and 1-min Apgar score are listed in Table 1. In addition, the occurrence of complications such as respiratory support, bronchopulmonary dysplasia (BPD), severe retinopathy of prematurity, and NEC after birth of preterm infants is also shown in Table 1.

Principal component analysis (PCA) was performed on the proteomic data of whole urine and urine supernatant of preterm infants, and the results showed that there was no obvious clustering of urinary proteins between the two groups (Fig. 1a). A total of 2,607 proteins were identified in the two groups, of which 2,469 overlapped proteins in the two group, 82 proteins were unique to the whole urine samples (Fig. 1b), and 56 proteins were unique to the supernatant samples (Fig. 1b).

The Gene Ontology (GO) knowledgebase was used to analyze the molecular functions and biological processes of the differentially expressed proteins. We found that among the top 10 differentially expressed proteins, the urine supernatant was mainly composed of protein fragments and extracellular secreted MST1 protein. In addition to cell debris and secreted extracellular protein MUC5B, there were mainly cytoplasmic protein, nuclear protein, and membrane protein in whole urine. Among the top 20 proteins with differential expression, 8 were decomposed protein fragments, and the remaining 12 proteins had molecular functions mainly including ribosomal proteins, proteases, intercellular adhesion proteins, and cytoskeletal proteins (Table 2). The differentially expressed proteins were mainly involved in the biological processes of rRNA processing, rRNA metabolic process, ncRNA processing, ribosome biogenesis, ncRNA metabolic process, RNA processing, translational initiation, ribonucleoprotein complex biogenesis maturation of SSU-rRNA (Fig. 2). Further analysis showed that, in addition to the detected protein fragments, among the top 10 proteins enriched in whole urine and urine supernatant of preterm infants, the main proteins enriched in whole urine were RPS16, PSMA6, VDAC1, NIT2, MUC5B, RNPEP, DHRS2, RPS16, PSMA6, VDAC1, NIT2, and MUC5B. The main proteins enriched in the urine supernatant after centrifugation were MST1, PHC3 and DSG3 (Fig. 3a).

By analyzing the cellular component (CC) of GO database, the subcellular localization of differentially expressed proteins in the two groups of urine samples was annotated and statistically analyzed. We found that the differentially expressed proteins were significantly concentrated in the cytoplasm, extracellular region, and membrane (Fig. 3b).

Based on the existing literature on neonatal urinary proteomics, we reviewed the literature to screen urinary protein markers related to neonatal diseases such as BPD and other respiratory diseases, necrotizing enterocolitis (NEC), acute kidney injury, and kidney development (Table 3). Ahmed et al. [19] found that the levels of CHI3L1, MMP-9, FZD6, CP, C1QC, and FKBP1A were related to the incidence or severity of BPD, which could be used as potential biomarkers for BPD screening. In a study of neonatal NEC, Sylvester et al. [21] identified seven urinary protein biomarkers (alpha-2-macroglobulin-like protein 1, cluster of differentiation protein 14, cystatin 3, fibrinogen alpha chain, pigment epithelium-derived factor, retinol-binding protein 4, and vasolin) can provide accurate diagnosis and prognosis information for neonates with suspected NEC. In a proteomic study of early urinary biomarkers in preterm infants with acute kidney injury, Jung et al. [31] found that annexin A5, neutrophil gelatinase-associated lipocalin, and protein S100P has important value in early and accurate prediction of acute kidney injury in preterm infants. Charlton et al. [32], in their study on the relationship between urinary protein and kidney maturation during kidney development in preterm infants, found that the level of insulin-like growth factor-binding protein-1, -2, and -6, monocyte chemotactic protein-1, CD14, and sialic acid-binding Ig-like lectin 5 can reflect kidney maturation. Starodubtseva et al. [22] found in their study on the difference of urinary proteome in preterm infants between infectious and noninfectious respiratory diseases that proteins involved in cell adhesion (CDH-2, -5, -11, NCAM1, TRY1, DSG2), metabolism (LAMP1, AGRN, TPP1, GPX3, APOD, CUBN, IDH1), regulation of enzyme activity (SERPINA4, VASN, GAPDH), inflammation and stress response (CD55, CD93, NGAL, HP, TNFR, LCN2, AGT, S100P, SERPINA1/C1/B1/F1) protein levels could significantly distinguish neonates with infectious respiratory diseases from those without infectious diseases. We further analyzed the enrichment of the selected urinary protein biomarkers related to neonatal diseases in whole urine and urine supernatant and found that there was no significant difference in the enrichment of urinary protein biomarkers between whole urine and urine supernatant in these disease states. It is suggested that whole urine and urine supernatant may have little effect on the diagnosis of neonatal diseases in the processing of urine samples.

With the improvement of neonatal intensive care, more and more preterm infants and very low and extremely low birth weight infants are born. More and more attention has been paid to not only the early survival after birth but also the long-term diseases of various systems [5, 33]. In recent years, more and more epidemiological data suggest that preterm birth is an important risk factor for long-term diseases of various systems such as the nervous system, respiratory system, and urinary system [6, 34‒36]. Our previous studies have found that preterm birth can lead to the reduction of renal podocyte number and further drive the deterioration of podocyte loss, which ultimately leads to the increased risk of long-term chronic kidney disease [37‒39]. Early monitoring, early diagnosis, and early treatment are important for reducing the risk of long-term chronic diseases caused by preterm birth.

At present, clinical and laboratory tests have a certain degree of invasive and hysteresis property in the diagnosis and treatment of the disease. Finding a noninvasive method for early detection and monitoring of disease is an important goal in the field of neonatology. In recent years, proteome analysis has been increasingly studied for disease biomarker research. Proteomic analysis has not been fully developed in clinical practice to predict disease. Noninvasive proteomics has attracted more and more attention for clinical disease diagnosis and prognosis monitoring. It has great potential in the field of disease mechanism research, early clinical diagnosis, efficacy evaluation, and new drug development [40, 41].

In many disease states, there will be related protein changes in the body’s blood circulation, which appear in the urine after glomerular filtration, especially the related protein changes in urinary system diseases can be reflected in the urine [42]. For preterm infants, urine is a valuable biological specimen that can be obtained noninvasively and monitored dynamically. At present, some studies have used urine proteomics to screen biomarkers in neonatal diseases such as bronchopulmonary dysplasia, NEC, and acute kidney injury [19‒21], in order to explore the pathogenesis of diagnosis and prevention of related diseases. Currently, there is no uniform preparation method for the processing of urine samples for urine proteomics research. There are two main methods to prepare urine specimens in the current study. One is the collection of whole urine specimens, and the other is the urine supernatant after removing the urine sediment [43, 44]. However, it has not been clarified whether there are differences in protein profiles and biomarker screening between the two urine samples [22, 23]. In this study, preterm birth was used as a disease model to investigate the differences in protein profiles between the whole urine and urine supernatant of preterm infants.

This study found that a total of 2,607 proteins were detected in the urine samples of preterm infants. There was no obvious clustering between the two groups of proteins by PCA, indicating that most of the urinary proteins in the two urine processing methods were common to the two groups. Among them, 82 proteins were unique to the whole urine samples and 56 proteins were unique to the supernatant samples. In this study, we further analyzed the differentially enriched proteins between the two groups and found that except for detectable protein fragments, the main enriched protein in the urine supernatant samples was hepatocyte growth factor-like protein encoded by MST1 gene. Previous studies have found that the gene MST1 is associated with the risk of inflammatory bowel disease [45]. Compared with the whole urine, MST1 is significantly enriched in the supernatant urine, which has important theoretical guidance for the prediction of neonatal inflammatory bowel disease and subsequent disease monitoring and prognosis. The main enriched proteins in whole urine samples were ribosomal proteins, intercellular adhesion proteins, and cytoskeletal proteins. Through the subcellular localization of differential proteins, it was found that the enriched differential proteins in the urine protein supernatant were located in the extracellular region and the protein enriched in whole urine was localized to the bladder, extracellular region, membrane. This study further analyzed the enrichment of urinary protein markers in whole urine and urine supernatant found in the existing literature in neonatal BPD and other respiratory diseases, NEC, acute kidney injury, and kidney development. We found no significant difference in the protein enrichment levels of these biomarkers between these two different urine processing methods. This suggests that the proteomics of whole urine and urine supernatant may have little effect on the diagnosis of the above neonatal diseases.

At present, the clinical research of predicting the occurrence of preterm birth related diseases by urine protein biomarkers is still in the exploratory stage. The existing studies can detect biomarkers of neonatal diseases through urine proteomics technology in neonatal BPD and other respiratory diseases, NEC, acute kidney injury, and kidney development, which provide an important theoretical basis for clinical disease diagnosis, disease progression, disease risk assessment, and clinical treatment outcome evaluation. However, in present various studies, there is no unified standard for the treatment of urine samples. When to use whole urine or urine supernatant for detection, and whether these two treatment methods have an impact on the screening of disease protein biomarkers is unknown. Therefore, this study analyzed the protein profiles in the whole urine and urine supernatant of premature infants. The aim is to elucidate the differences of two different treatment methods in urine proteomics of the same premature infant and to provide theoretical basis for sample treatment methods for future research. This study has some possible limitations because of the small sample size, and further large-scale studies are needed to confirm the clinical utility. In addition, LC-MS/MS technique itself has a high sensitivity and can detect low abundance of proteins. Thus, after detected by LC-MS/MS, further identification in urine samples by different methods such as ELISA should also be important, which did not performed in present study. More significantly, in future studies, we can track the urine protein profile of preterm infants at different life stages through longitudinal studies and include a more diverse population of preterm infants such as gestational age and birth weight to gain insights into how various factors influence urine protein profiles.

In conclusion, this study provides a theoretical basis for the preparation of urine samples for the prediction of different neonatal diseases by comparing the proteomics differences between the whole urine and urine supernatant of preterm infants. Different urine preparation methods have different enrichment degrees for different subcellular localization proteins. Based on the current research into urinary protein biomarkers associated with neonatal diseases, our study revealed that there was no statistically significant difference in the enrichment of relevant urinary protein biomarkers between whole urine and urine supernatant. This indicates that both whole urine and urine supernatant are suitable for investigating urinary protein biomarkers related to neonatal conditions among these proteins. However, it is crucial to carefully consider the urine collection methods when analyzing urine sample-specific proteins which identified in present study.

We acknowledge support from Tianjin Key Medical Discipline (Specialty) Construction Project.

This study protocol was reviewed and approved by Tianjin Central Hospital of Gynecology and Obstetrics Institutional Review Board, Approval No. 2020KY049. All study protocols adhered to the Declaration of Helsinki, and written informed consent was obtained from all parents of the participants before their inclusion in this study.

All the authors declared no competing interests.

L.Z. is sponsored by Tianjin Health Research Project (Grant No. TJWJ2022QN087) and Open Fund of Tianjin Central Hospital of Gynecology Obstetrics/Tianjin Key Laboratory of Human Development and Reproductive Regulation (Grant No. 2022XH06). F.D. is sponsored by Tianjin Health Commission (Grant No. TJWJ2021QN054), Tianjin Science and Technology Committee (21JCQNJC01650), and China International Medical Foundation (Grant No. Z-2019-41-2101-04).

F.D. and J.Z. designed the study; L.Z., J.Z., X.W., and F.D. carried out experiments; L.Z., D.W., and F.D. analyzed the data; L.Z. and F.D. made the figures; L.Z., J.Z., and F.D. drafted and revised the manuscript; and all authors approved the final version of the manuscript.

Lulu Zhang and Xueyan Wang contributed equally to this work.

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (https://proteomecentral.proteomexchange.org) via the iProX partner repository with the dataset identifier PXD048786. The mass spectrometry proteomics data will be made available through iProX upon acceptance of the manuscript for publication. The data that support the findings of this study are available from the corresponding author ([email protected]) upon reasonable request.