Introduction. Kidney transplantation is the preferred treatment for patients with end-stage kidney disease. Although it increases life expectancy and improves patients' quality of life, long-term survival of allografts is hampered by rejection, which is the primary cause of kidney failure. Existing biomarkers of kidney function (serum creatinine, estimated glomerular filtration rate) and injury (proteinuria) lack the specificity and sensitivity to accurately detect allograft rejection and other causes of injury. Our research therefore aimed to identify novel noninvasive biomarkers of kidney allograft injury, focusing on extracellular vesicles (EVs) containing specific DNA cargo (vDNA) and cell-free DNA (cfDNA) in the urine of kidney allograft recipients. We aimed to determine whether the concentration and size of EVs and the quantity of vDNA and cfDNA present were related to allograft injury.
Hypotheses. 1. A pure population of extracellular vesicles can be isolated from the urine of kidney transplant recipients, the concentration and size of which are related to the histopathologic changes of the kidney allograft biopsy, providing insight into kidney allograft injury even in the presence of stable kidney function. 2. From the urine of kidney transplant recipients, we can extract the cell-free DNA and extracellular vesicle-bound DNA released from the donor's cells, the quantity of which is related to the histopathologic changes observed in the kidney allograft biopsy.
Methods. We enrolled 93 adult kidney transplant recipients in a prospective clinical study, conducted in two parts. We examined EVs and their DNA cargo, which we compared to the cfDNA at the time of biopsy in a prospective cohort study (n = 40). We conducted a prospective longitudinal study (n = 53) to observe the characteristics of vDNA at multiple time points within the 1st year after transplantation. We categorized participants into three groups based on the presence and phenotype of allograft injury: normal histology (NH), rejection injury (RI), and non-rejection injury (NRI). We extracted genomic DNA from biopsy and blood samples to genotype donor-recipient pairs based on single nucleotide polymorphism using polymerase chain reaction (PCR). We optimized a protocol using size-exclusion chromatography to isolate EVs from second morning urine. We then examined the characteristics of EVs using immunolabeling, Western blotting, nanoparticle tracking analysis, and scanning electron microscopy. We extracted cfDNA from the urine and vDNA from EVs. Based on the identified genotypes, we determined the quantity and other characteristics of v/cfDNA (yield, degree of fragmentation, total and donor-derived DNA copy number, and donor-derived fraction) using digital PCR. To account for variability in urine concentration, certain characteristics were normalised to urine creatinine levels.
Results. Using our optimized protocol, we have successfully isolated a pure population of EVs from the urine of kidney transplant recipients. In the prospective cohort study patients from both groups with allograft injury had significantly larger EVs (P < 0.05), while we found no differences in the normalized EV concentration between the patient groups. Normalized yield, DNA copy number and donor-derived fraction of both DNA entities were significantly higher in the RI and NRI groups (P < 0.05). CfDNA was significantly more fragmented than vDNA. The analyzed characteristics of vDNA and cfDNA correlated significantly with the Banff scores of the histopathological lesions. Longitudinal follow-up of patients showed that after 1 month, the donor-derived vDNA fraction significantly increased, while normalized vDNA and donor-derived vDNA copy number significantly decreased (P ⡤ 0.001). A vDNA copy number greater than 9,027 at month 6 predicted the occurrence of kidney allograft injury one year after transplantation (area under the curve 0.730, 81.8% sensitivity, 73.3% specificity; P = 0.025).
Conclusion. We were the first to study the DNA cargo of EVs from the urine of kidney transplant recipients and the first to show a correlation between the quantity of vDNA/cfDNA and histologic lesions. We showed that EV size reflects allograft injury, whereas EV concentration, in contrast, is not related to injury phenotype. Thus, we partially confirmed the first hypothesis. We confirmed the second hypothesis, as the quantity of donor-derived vDNA and cfDNA in the urine, is associated with the histopathologic lesions in the allograft.
|
