Relationships among injury, fibrosis, and time in human kidney transplants

Abstract

Background: Kidney transplant biopsies offer an opportunity to understand the pathogenesis of organ fibrosis. We studied the relationships between the time of biopsy after transplant (TxBx), histologic fibrosis, diseases, and transcript expression.

Methods: Expression microarrays from 681 kidney transplant indication biopsies taken either early (n = 282, <1 year) or late (n = 399, >1 year) after transplant were used to analyze the molecular landscape of fibrosis in relationship to histologic fibrosis and diseases.

Results: Fibrosis was absent at transplantation but was present in some early biopsies by 4 months after transplant, apparently as a self-limited response to donation implantation injury not associated with progression to failure. The molecular phenotype of early biopsies represented the time sequence of the response to wounding: immediate expression of acute kidney injury transcripts, followed by fibrillar collagen transcripts after several weeks, then by the appearance of immunoglobulin and mast cell transcripts after several months as fibrosis appeared. Fibrosis in late biopsies correlated with injury, fibrillar collagen, immunoglobulin, and mast cell transcripts, but these were independent of time. Pathway analysis revealed epithelial response-to-wounding pathways such as Wnt/β-catenin.

Conclusion: Fibrosis in late biopsies had different associations because many kidneys had potentially progressive diseases and subsequently failed. Molecular correlations with fibrosis in late biopsies were independent of time, probably because ongoing injury obscured the response-to-wounding time sequence. The results indicate that fibrosis in kidney transplants is driven by nephron injury and that progression to failure reflects continuing injury, not autonomous fibrogenesis.

Trial registration: INTERCOM study (www.clinicalTrials.gov; NCT01299168).

Funding: Canada Foundation for Innovation and Genome Canada.

Figures

Figure 7. The nephron-centric model of renal transplant fibrosis based on the events in the first year after transplant as reflected in early indication biopsies.

This model reflects the findings in the present study, placing atrophy-fibrosis in the general context of response to wounding and particularly in relationship to the nephron decision to either restore function or shut down. The model assumes that many factors can influence the nephron decision, including severity of wounding, biological aging and somatic cell senescence, and podocyte numbers, with tubuloglomerular feedback as an effector-controlling filtration. The model postulates that similar changes occur in late kidneys with injury due to diseases or other stresses but that the time series is obscured by the ongoing injuries. AKI, acute kidney injury.

Figure 6. Expression of the top 30 TxBx-corrected transcripts associated with high fibrosis across the panel of cultured cells compared with normal kidney control.

(A) Hierarchical clustering of the top increased transcripts. (B) The top decreased transcripts. Rows represent z-scores of transcripts, and columns represent cell types. Hierarchical clustering was done on rows and columns using Euclidean distance as the similarity measure with a full linkage. Human renal tubular epithelial cells, RPTEC; human umbilical vein endothelial cells, HUVEC.

Figure 5. The dependence of the global transcript landscape of atrophy-fibrosis on time after transplant (TxBx).

The association strength of all interquartile range–filtered (IQR-filtered) probe sets to atrophy fibrosis (ci score >1) as tested by the Bayes t test. Probe sets belonging to the following transcript set were labeled as follows: mast cell transcripts (MCATs), red circles; immunoglobulin transcripts (IGTs), blue circles; acute kidney injury (AKI) transcripts top 30, green circles; all 950 AKI transcripts (IRRAT950), small green circles; fibrillar collagen (FICOL) transcripts, yellow circles; and remaining IQR filtered probe sets, gray dots. Each symbol represents an individual probe set. (A) biopsies taken before 1 year TxBx (n = 282). (B) Biopsies taken after 1 year TxBx (n = 399). (C) All biopsies (n = 681). (D) All biopsies with TxBx correction (n = 681).

Figure 4. The global transcript landscape of atrophy fibrosis in the discovery and validation sets.

The association strength of all IQR-filtered probe sets to atrophy fibrosis (ci score >1) is shown as P values from the ci score >1 Bayes t test (x axis), with the corresponding fold changes (y axis). We visualized probe sets belonging to the following transcript sets: mast cell (MCATs), red; immunoglobulin (IGTs), blue; acute kidney injury (AKI), green; and fibrillar collagen (FICOL), yellow; and remaining interquartile range–filtered (IQR-filtered) probe sets, gray dots. Each symbol represents an individual probe set. (A) Discovery set biopsies only (n = 401). (B) Validation set biopsies only (n = 280). (C and D) Relationship of P values (Spearman correlation coefficient = 0.51) (C) and fold changes (D) (Spearman correlation coefficient = 0.71) between the discovery and validation sets.

Figure 3. Moving averages for atrophy fibrosis–related features vs.

time after transplant. The left y axis represents average molecular scores (immunoglobulin [IGT], mast cell [MCAT], acute kidney injury [AKI], and fibrillar collagen [FICOL] transcripts), and the right y axis average histologic ci scores (fibrosis). Biopsies are ordered by time after transplant and then the means (of both the x and y axis values) are plotted based on a sliding window of size 100 biopsies.

Figure 2. Death-censored transplant survival.

(A and B) Death-censored transplant survival in whole patient population (A) and in population of patients with fibrosis (B). A random biopsy per patient was selected. Bx, biopsy; n, number of biopsies; F, number of failures.

Figure 1. Relationship between the histologic scores for fibrosis and time after transplant to biopsy.

Distribution of fibrosis lesions (ci scores) in time intervals after transplant. The block height represents the number of biopsies with each ci score in each time interval. d, days after transplant; mo, months after transplant; yr, years after transplant; N, number of biopsies in each time interval; and %ci>1, percentage of biopsies with ci score >1 in each time interval.