Circulating T cell specific extracellular vesicle profiles in cardiac allograft acute cellular rejection

Abstract

There is a critical need for biomarkers of acute cellular rejection (ACR) in organ transplantation. We hypothesized that ACR leads to changes in donor-reactive T cell small extracellular vesicle (sEV) profiles in transplant recipient circulation that match the kinetics of alloreactive T cell activation. In rodent heart transplantation, circulating T cell sEV quantities (P < .0001) and their protein and mRNA cargoes showed time-specific expression of alloreactive and regulatory markers heralding early ACR in allogeneic transplant recipients but not in syngeneic transplant recipients. Next generation sequencing of their microRNA cargoes identified novel candidate biomarkers of ACR, which were validated by stem loop quantitative reverse transcription polymerase chain reaction (n = 10). Circulating T cell sEVs enriched from allogeneic transplant recipients mediated targeted cytotoxicity of donor cardiomyocytes by apoptosis assay (P < .0001). Translation of the concept and EV methodologies to clinical heart transplantation demonstrated similar upregulation of circulating T cell sEV profiles at time points of grade 2 ACR (n = 3 patients). Furthermore, T cell receptor sequencing of T cell sEV mRNA cargo demonstrated expression of T cell clones with intact complementarity determining region 3 signals. These data support the diagnostic potential of T cell sEVs as noninvasive biomarker of ACR and suggest their potential functional roles.

1. Introduction

Allograft surveillance for acute cellular rejection (ACR) mediated by recipient T cells remains a critical component of patient care, as acute rejection and immunosuppression-related complications remain principal drivers of patient morbidity and mortality.^1,2 Currently there is a critical need for development of a biomarker platform that dynamically provides a noninvasive physiological window into the conditional status of the recipient T cell alloreactive state. Peripheral blood white blood cell counts, T cell counts, and the phenotypes of their subclasses are not significantly altered during ACR.^3,4 Assays that analyze the inflammatory profile of peripheral blood mononuclear cells during ACR lack adequate diagnostic accuracy and are not T cell specific.³ Therefore, a biomarker platform that would enable simple, repeatable, reliable read-out of the immune synapse between the donor tissue and recipient alloreactive T cells has potential to provide a novel mechanistic and diagnostic window into ACR.

Small extracellular vesicles (sEVs) are produced by many cell types and released into bodily fluids, including blood.^5–7 Their membranes contain surface markers of their parent cell type, and their contents reflect the conditional changes imposed on their cells of origin. sEVs are also implicated in immune regulatory processes,^8–11 but their physiological role in transplant rejection and their diagnostic potential as biomarkers of ACR need better understanding. Previously, we investigated a novel circulating donor tissue-specific sEV platform that enabled noninvasive monitoring for acute rejection with high accuracy,^12–16 demonstrating that circulating donor tissue sEVs and their cargo profiles herald early ACR in a time-specific manner. Herein, we hypothesize that ACR-associated activation of alloreactive T cells would lead to time-specific changes in circulating T cell sEV quantities and their intra-EV protein and RNA cargoes. Furthermore, given the immunoregulatory potential of EVs, we propose that T cell sEVs play a functional role in mediating the pathophysiology of ACR itself. Herein, we describe development of methodology for T cell sEV enrichment from peripheral blood, report our investigation of this hypothesis in a rodent heart transplantation model of ACR, and validate its methodological translation to clinical heart transplantation.

2. Materials and methods

Supplementary information may be found in Supporting Information section.

2.1. Animal care

All experiments were conducted in accordance with approved protocols through the University of Pennsylvania (Protocol 805005).

2.2. Study design

A schematic of the study design is shown in Figure 1A. See Supporting Information.

Figure 1.

Study design and development of T cell sEV platform. (A) Three study arms designed to ascertain circulating T cell sEV profiles in a heterotopic heart transplantation model is shown. In the Maintenance arm, a BALB/c heart was transplanted into a T cell and B cell immunodeficient C57BL/6 Prkdc^SCID recipient to assess for nonspecific CD3 sEV signal and donor passenger T cell sEV signal in recipient blood. In the Rejection arm, a BALB/c [H2-k^d] heart was transplanted into a wild type C57BL/6 [H2-k^b] recipient across full MHC mismatch, which reliably led to ACR mediated allograft asystole by day 12 to 14. In the Control arm, syngeneic C57BL/6 into C57BL/6 transplant was performed to assess for changes in circulating T cell sEV signal from isograft transplantation. (B) To assess if T cells may release sEVs with specific markers, first, Jurkat T cells were studied. NTA (scatter mode) of sEVs from Jurkat T cell supernatant is shown, along with mean sEV size, quantity, and concentration of particles per μg of protein (n = 10). (C) Wide field electron microscopy of Jurkat T cell sEVs is shown. (D) Western blot of Jurkat sEVs showed expression of exosome markers TSG101, CD63, flotillin-1, and T cell markers CD4, CD8, and TCR. Jurkat sEVs were also checked for absence of expression of calnexin and cytochrome c per MISEV guidelines. (E) NTA images of isolated plasma sEVs from posttransplant time points in all 3 study arms are shown, along with mean sEV size and quantity (# particles/μg sEV protein). (F) Electron microscopy of plasma sEVs from day 7 posttransplant in the Control, Rejection, and Maintenance arms is shown. (G) Before downstream analysis, western blotting of plasma sEVs was performed for expression of exosome markers CD63, flotillin-1, and TSG101, with absence of cytochrome c, calnexin, and apolipoprotein E, per MISEV guidelines. (H) CD3-expressing T cell sEV subset signal (red) compared to the total plasma sEV pool (blue) on the NTA fluorescence mode is shown for animal receiving syngeneic transplant (Control arm). Percentage nanoparticles positive for CD3 expression is also shown. (I) Circulating CD3-expressing sEV signal over serial time points in one animal undergoing heart transplantation in the Maintenance arm is shown. Naive wild type C57BL/6 plasma sEVs and IgG isotype control are also shown. (J) CD3 sEV signal in total plasma sEV pool from a single recipient animal over serial follow-up after full MHC mismatch heart transplant is shown (Rejection arm), along with naive, wild type control and IgG isotype control. Anti-CD3ε antibody-conjugated quantum dot was utilized for T cell sEV subset quantitation on the NTA fluorescence mode. Time-specific changes in the T cell sEV signal were seen only in the Rejection arm recipient animal. Abbreviations: EV, extracellular vesicle, MISEV, Minimal Information for Studies of Extracellular Vesicles; MHC, major histocompatibility complex; NTA, nanoparticle tracking analysis; sEV, small extracellular vesicle; TCR, T cell receptor.

2.3. Heterotopic heart transplantation and posttransplant monitoring

The heterotopic heart transplantation procedure was previously described.⁵ See Supporting Information.

2.4. Mouse plasma sEV isolation

sEVs were isolated from 250 μL to 300 μL of C57BL/6 mouse and BALB/c mouse plasma samples using size exclusion chromatography along with ultracentrifugation.^13–15 Based on recommendations per the Minimal Information for Studies of Extracellular Vesicles (MISEV) 2018 position statement,¹⁷ EVs investigated in this study are categorized as small EVs <200 nm (sEVs). In the manuscript, the distinct subpopulation of CD3-bearing sEVs enriched based on anti-CD3 antibody bead incubation with whole plasma sEVs is referred to as “T cell sEVs.” See Supporting Information.

2.5. Antibodies

Antibodies used for sEV protein analysis are shown in Table 1.

Table 1.

List of antibodies utilized in this study is shown.

Antibody	Vendor	Catalog number	Dilution
Perforin 1	Santa Cruz Biotechnology	SC-373943	1:200
CD38 (H-11)	Santa Cruz Biotechnology	SC-374650	1:200
Flotillin-1	Santa Cruz Biotechnology	SC-74566	1:200
TSG 101 (C-2)	Santa Cruz Biotechnology	SC-7964	1:200
TSG101	Proteintech	28283-1-AP	1:500
FasL (D1N5E)	Cell Signaling Technology	68405	1:500
Goat anti-Rabbit IgG-HRP	Santa Cruz Biotechnology	SC-2004	1:5000
Goat-anti-Mouse IgG-HRP	Santa Cruz Biotechnology	SC-2005	1:5000
CD3	Santa Cruz Biotechnology	SC-1179	1:200
TCR	Santa Cruz Biotechnology	SC-65737	1:200
CD4	Santa Cruz Biotechnology	SC-19641	1:200
CD3	Santa Cruz Biotechnology	SC-20080	1:200
Granzyme B	Santa Cruz Biotechnology	SC-8022	1:200
IFNy	Proteintech	15365-1-AP	1:500
FOXP3	Proteintech	22228-1-AP	1:500
Flotillin 1	Proteintech	15571-1-AP	1:500
MHC Class II	Abcam	Ab55152	1:500
CD8	Santa Cruz Biotechnology	SC-1177	1:200
SERCA2	Santa Cruz Biotechnology	SC-73022	1:200
CD63	Santa Cruz Biotechnology	SC-365604	1:200
Cytochrome c	Santa Cruz Biotechnology	SC-65396	1:200
H2-Kd (SF1-1.1)	Santa Cruz Biotechnology	SC-53852	1:1000
CD3	Biolegend	100203	1:400
CD19	Biolegend	159807	1:400
B220	Biolegend	103207	1:400
CD4	Biolegend	100431	1:400
CD8	Biolegend	140408	1:400

2.6. Nanoparticle detector tracking analysis

sEVs were analyzed on the NanoSight NS300 nanoparticle detector on the light scatter mode for quantification and size distribution (Malvern Instruments). Experimental sample analysis was performed using NTA3.4 software. T cell specific surface marker detection on sEVs was performed using the fluorescence mode. Secondary antibody conjugated to quantum dot with emission at 605 nm was utilized for fluorescence detection of T cell marker CD3 as described previously.^13,14

T cell sEV signal was quantified as:

$$\frac{\text{CD3}\text{fluorescence}}{\text{CD3}\text{light}\text{scatter}}-\frac{\text{IgG}\text{isotype}\text{fluorescence}}{\text{IgG}\text{isotype}\text{light}\text{scatter}}$$

2.7. RNA and protein cargo analysis

RNA was extracted from T cell sEVs using Trizol reagent per manufacturer protocol (Qiagen). For protein isolation, sEVs were lysed in 1× radioimmunoprecipitation assay buffer with 1× protease inhibitor cocktail (Sigma-Aldrich Co). Protein concentration was measured using bicinchoninic acid assay. See Supporting Information.

2.8. Reverse transcription PCR (RT-PCR)

Total RNA (25–50 ng) was reverse transcribed with the SuperScript III one-step RT-PCR system (Life Technologies) for gene expression validation per manufacturer protocol. Primers used are shown in Table 2.

Table 2.

RT-PCR primer sequences used in this study are shown.

Gene	Forward	Reverse
CD3ε	AAGTCGAGGACAGTGGCTACTAC	CATCAGCAAGCCCAGAGTGATACA
IFN-γ	TCAAGTGGCATAGATGTGGAAGAA	TGGCTCTGCAGGATTTTCATG
CXCL10	GGATGGCTGTCCTAGCTCTG	ATAACCCCTTGGGAAGATGG
FoxP3	CAGCTGCCTACAGTGCCCCTAG	CATTTGCCAGCAGTGGGEAG
IL2	CCTGAGCAGGATGGAGAATTACA	TCCAGAACATGCCGCAGAG
β-actin	GGCTGTATTCCCCTCCATCG	CCAGTTGGTAACAATGCCATGT
TCR	GTGGGTGAATGGCAAGGAGGTCCAC	GGTTTGGGTGAGCCCTCTGGC

2.9. Quantitative reverse transcription PCR (RT-qPCR)

Synthesis and amplification of cDNA was performed using TaqMan Advanced miRNA cDNA synthesis kit (A28007) (Applied Biosystems). Ct values were analyzed by the 2-(ddCt) method.¹⁸ See Supporting Information.

2.10. T cell sEV miRNA cargo analysis

miRNA reads were processed and analyzed using the GeneGlobe QIAseq miRNA Library Kit – Primary Quantification (Qiagen). Differential miRNA expression for day 4 and day 7 sEVs (n = 6) was computed using edgeR (version 3.30.3) to determine the fold change (fc) and P value. Significantly enriched apoptotic or immune-related pathways were then selected by false discovery rates <0.05 using REVIGO.¹⁹ See Supporting Information.

2.11. Fluorescence activated cell sorting

Blood samples were collected using ethylenediaminetetraacetic acid tubes and subsequently stained with the following anti-mouse cell surface markers: CD3 (BioLegend), CD4 (BioLegend), CD8 (BioLegend), CD19 (BD Pharmingen), and B220 (BioLegend). For each sample, 10 μL of whole blood was incubated with a master mix containing the antibodies for 20 minutes at room temperature in the dark. Samples were analyzed on BD LSR Fortessa flow cytometer (BD Biosciences), first gated using forward scatter area (FSC-A) by side scatter area (SSC-A), then single cells using FSC-A by forward scatter height (FSC-H). This population was divided into B220⁺/CD19⁺ cells and CD3⁺ cells divided into CD4⁺/CD8⁺ cells. The percentage of the parent population was analyzed using FlowJo software.

2.12. Cardiomyocyte isolation and apoptosis assay

Primary cardiomyocytes were isolated from whole heart preparations as previously described.^20,21 Apoptosis assay was performed using Click-iT Plus TUNEL assay for in situ apoptosis detection, using Alexa Fluor 647 dye kit (C10619, ThermoFisher) following the manufacturer’s suggested protocol. See Supporting Information.

2.13. Human plasma sEV characterization

University of Pennsylvania Institutional Review Board (Protocol 831237) approval was obtained. From 4 heart transplant patients, blood was collected at times of surveillance endomyocardial biopsy (EMB) up to 100 days postoperatively. sEV characterization was performed as previously described.^13,16

3. Results

3.1. Jurkat T cells release sEVs carrying T cell surface markers in vitro

Jurkat T cells were cultured in EV-free media, and supernatant was collected for sEV isolation. sEV phenotype was characterized by nanoparticle tracking analysis (NTA) (Fig. 1B, Supplementary File 1), electron microscopy (Fig. 1C), and protein marker expression (Fig. 1D) per MISEV guidelines.¹⁷ On western blot, in addition to exosome marker expression (flotillin 1, TSG101, CD63), T cell markers CD4, CD8, and T cell receptor (TCR) were also detected (Fig. 1D).

3.2. ACR leads to time-specific changes in circulating T cell sEV quantities

Plasma sEVs were isolated at several time points in the 3 study arms and quality checked for size distribution on NTA (Fig. 1E, Supplementary Files 2–4), electron microscopy (Fig. 1F), and protein expression (Fig. 1G) per MISEV guidelines.¹⁷ Based on these parameters, we previously demonstrated that whole plasma sEV profiles are similar in the Rejection versus Maintenance arms.¹⁴ The T cell sEV subpopulation was quantified within the total plasma sEV pool using anti-CD3 antibody-conjugated quantum dots on NTA fluorescence mode. For each study arm, serial blood samples from single animals were analyzed over several time points to assess if CD3 sEV signal varied with ACR. In the Control (Fig. 1H) and Maintenance arms (Fig. 1I), CD3 sEV signal remained near baseline levels over follow-up. In the Rejection arm, T cell sEV signal increased by day 4, peaked around day 7 to 8, and decreased to baseline levels by day 15 posttransplant when allograft rejection was complete (Fig. 1J). Based on these findings, we performed more heart transplants to characterize T cell sEVs during ACR.

A scatter plot display of circulating T cell sEV signal for the Maintenance, Control, and Rejection arms is shown in Figure 2A. This demonstrated that T cell sEV quantities markedly increased by postoperative days (PODs) 4 to 5, peaked on POD 7, and returned to baseline levels by POD 15 in the Rejection arm only. By 2-way analysis of variance (ANOVA) comparing the study arms, CD3 sEV signal was significantly altered in the POD 4 versus 7 samples in the Rejection arm compared to the Control arm (P <.0001) and Maintenance arm (P <.0001). However, at these time points, there was no significant difference in the CD3 sEV signal between the Control arm and Maintenance arm (P = .07). Considering the possibility that this signal change with ACR may just be reflecting concomitant changes in circulating T cell numbers, we performed fluorescence activated cell sorting (FACS) analysis of peripheral blood mononuclear cells in the Control and Rejection arms for quantitative expression of CD3⁺ T cells and their helper (CD4) and cytotoxic (CD8) subpopulations (Fig. 2B, C). CD3⁺ T cell quantities or the distribution of CD4⁺ and CD8⁺ T cell subpopulations were not significantly altered in the peripheral blood at the day 0 versus day 7 posttransplant time points (Fig. 2C). By 2-way repeated measures ANOVA, CD3 T cells (P = .27), CD4 T cells (P =.49), and CD8 T cells (P =.11) were not significantly altered between the Rejection and Control arms (Fig. 2D). Taken together, this suggests that ACR leads to time-specific changes in circulating T cell sEV signal, without significant alterations in the peripheral blood CD3 T cell pool.

Figure 2.

Circulating T cell sEV signal heralds early ACR. (A) Scatter plot of circulating T cell sEV quantitative signal in the total plasma sEV pool is shown for the Maintenance (black), Control (blue), and Rejection (red) study arms. Each data point represents a sacrificed animal for the respective postoperative time point. In this ACR model, T cell infiltration is typically seen on POD 5, and donor heart rejection is complete (allograft asystole) by POD 12 to 14. With ACR seen only in the Rejection arm, time-specific increase in the T cell sEV signal was noted by POD 4/5 that peaked and persisted on PODs 7 and 9 and returned to baseline levels upon completion of allograft rejection. This time-specific CD3 sEV signal change was not noted in the Control and Maintenance arms. By 2-way ANOVA, CD3 sEV signal was significantly different in the Rejection arm compared to the Control arm (P <.0001) and the Maintenance arm (P <.0001) in a time-dependent manner (POD 4 versus 7). There was no significant difference in the CD3 sEV signal between the Maintenance versus Control arms (P = .07). (B) In the Control and Rejection arms, peripheral blood B cells (CD19⁺ B220⁺) and CD3⁺ T cell counts and their subpopulations, CD3⁺ CD4⁺ T cells and CD3⁺ CD8⁺ T cells, were analyzed by FACS to assess whether the changes in circulating T cell sEV signal in the Rejection arm reflected concomitant changes in circulating T cell numbers during ACR. Day 0 versus day 7 samples were analyzed as peak CD3 sEV signal was noted on day 7. Gating strategy for FACS is shown for 1 of 5 experiments in each study arm. In 3 animals, Sham operation was performed, where the animal’s abdomen was opened and the aorta and inferior vena cava were clamped and then released, and then the abdomen was closed. The animal was then sacrificed to assess for changes in circulating T cell numbers. (C) Representative FACS results for the Control arm, Rejection arm, and Sham operation are shown. The numbers of peripheral blood B cell and T cells and their CD4⁺ and CD8⁺ T cell subpopulations were similar in the Control versus Rejection study arms. (D) Scatter plot analysis of the results of FACS analysis for peripheral blood B cells and T cells is shown for animals sacrificed on POD 0 (4 hours after transplant) versus POD 7 (n = 5 animals in each study arm for each time point). On day 7, when the animal was sacrificed, peripheral blood (n = 5) and IVC blood near the heterotopically transplanted heart (n = 5) were analyzed by FACS. By 2-way repeated measures ANOVA, the number of circulating CD3⁺ T cells and their CD4 and CD8 T cell subpopulations were similar between the Rejection and Control arms on POD 0 and POD 7. Abbreviations: ACR, acute cellular rejection; ANOVA, analysis of variance; EV, extracellular vesicle; FACS, fluorescence activated cell sorting; IVC, inferior vena cava; POD, postoperative day; sEV, small extracellular vesicle.

3.3. Circulating T cell sEV cargoes are altered during ACR

Having demonstrated quantitative changes in T cell sEVs with ACR, we assessed whether their intra-EV cargoes are also altered during ACR. We analyzed whether T cell sEVs carry immunoregulatory markers typically associated with ACR. As T cell sEV levels peaked from baseline values between days 4 to 7, we performed cargo analysis at these 2 time points. From total plasma sEVs, T cell sEVs were purified using antibody-conjugated bead technology (Fig. 3A). The unbound sEV fraction was analyzed for absence of T cell markers CD3, CD4, and CD8 by western blot (Fig. 3B). To confirm that intact sEVs were successfully enriched, bead-bound sEVs were eluted and checked on NTA for detection of nanoparticles in the sEV range (Fig. 3C, Supplementary Files 5 and 6). Anti-CD3 antibody bead-bound sEVs showed expression of CD4, TCR, and immunoregulatory markers, including IFNγ, FoxP3, and Serca2 (Fig. 3D). These markers were upregulated in day 7 compared to day 4 samples in the Rejection arm but remained unchanged at low levels in the Control arm. T cell sEV mRNA cargoes also showed increased expression of CD3ε, TCR, IFN, and IL2 by RT-PCR on the day 7 time point in the Rejection arm but not Control arm (Fig. 3E). Collectively, this suggests that T cell sEVs carry immunoregulatory protein and mRNA markers, and ACR leads to changes in the expression of these markers.

Figure 3.

ACR leads to changes in T cell sEV cargoes, reflecting an alloreactivity profile. (A) Schematic for purification of circulating T cell sEVs is shown. Given the noted time course of T cell sEV signal during ACR, characterization of T cell sEV cargoes from days 4 and 7 after transplantation was performed. (B) Western blot of anti-CD3 antibody bead unbound sEV fraction was performed for confirmation of absence of CD3, CD4, and CD8 proteins as quality control. Jurkat cell positive control is also shown. Representative 1 of 3 experiments is shown. (C) Anti-CD3 antibody bead-bound sEVs were eluted and analyzed on NTA to confirm that intact nanoparticles in the sEV size range were being enriched using this methodology. Typically, when starting with 1×10¹⁰ sEVs as starting material, the resulting eluted bead-bound sEVs representing the enriched T cell sEV fraction yielded ~1 to 1.5 ×10⁸ sEVs, suggesting that the T cell sEV signal was being enriched between 66- to 100-fold compared to total plasma sEV analysis. Summary data for 5 independent experiments performed for the Rejection arm POD 7 is shown. (D) Western blot analysis of T cell sEV protein cargoes revealed upregulation of several T cell activation markers including IFNγ, Serca2, CD4, and TCR. Upregulation of these markers was noted in day 7 versus day 4 samples in the Rejection arm but not the Control arm. FoxP3 expression was also noted, indicating a regulatory T cell contribution to circulating T cell sEV subpopulation. Plasma sEVs from naive, wild type C57BL/6 are also shown to confirm that at the baseline, pretransplant setting, T cell markers are not readily detectable in plasma sEVs. Flotillin-1 is an exosome marker. One of 3 experiments is shown. (E) RT-PCR of T cell sEV mRNA showed upregulation of T cell activation markers in day 7 versus 4 samples in the Rejection arm but not the Control arm. CD3ε, TCR, IFNγ, and IL2 expression along with β-actin control is shown (1 of 3 experiments). Ab, antibody; ACR, acute cellular rejection; EV, extracellular vesicle; NTA, nanoparticle tracking analysis; RT-PCR, reverse transcription polymerase chain reaction; sEV, small extracellular vesicle; TCR, T cell receptor; WT, wild type.

3.4. T cell sEV miRNA cargo profiles

We performed next generation sequencing (NGS) of miRNA cargoes of circulating T cell sEVs to: (1) assess for candidate biomarkers of ACR and (2) provide mechanistic insights into potential functional roles of T cell sEVs. T cell sEV RNA cargo from 5 animals in the Rejection arm was pooled for single NGS analysis, with 3 such separate experiments performed for days 4 (n = 15) and 7 (n = 15). Principal component analysis showed distinction between days 4 versus 7, with one pooled outlier (Fig. 4A). We performed differential expression analysis of T cell sEV cargo (Fig. 4B) and identified the top 31 differentially expressed miRNAs in days 4 versus 7 (Fig. 4C). To determine their biological significance, we identified experimentally validated miRNA gene targets with miRTarBase and performed pathway analysis, which showed changes in immune regulatory processes in accordance with the temporal flow of the ACR process (Fig. 4D). Day 4 miRNA cargoes, representing initial stages of ACR, showed selective targeting of immune pathways involved in regulation of B cell activation, T cell differentiation, adaptive immune response, lymphoid organ development, and lymphocyte activation (Fig. 4D, E). Selective upregulation of apoptotic pathways was noted for day 7: regulation of programmed cell death, cytochrome c release from mitochondria, apoptotic process, and positive regulation of cell death (Fig. 4D, E), consistent with donor cytotoxicity associated with ACR. Next, we assessed the major miRNA targets implicated in these processes. Gene targets with ≥2 miRNA hits from the 31 differentially expressed miRNAs are shown in Figure 4F. Several of these gene targets are implicated in apoptosis pathways, especially HDAC4, PTEN, BCL2, IL-6, MEF2c, and SMAD7. Taken together, miRNA cargo pathway analysis suggested that T cell sEVs may play a pathophysiologic role in facilitating ACR.

Figure 4.

T cell sEV miRNA cargo profiles reflect alloactivation state. (A) Principal component analysis plot of Rejection arm day 4 and day 7 CD3 sEV NGS samples showed distinct miRNA cargo populations. (B) Volcano plot of fold change versus P value revealed differentially expressed miRNAs from day 4 to day 7. miRNAs upregulated in day 4 and 7 are highlighted in blue and orange, respectively. (C) Heatmap of top 31 differentially expressed miRNAs (|fc| > 2, P < .15) is shown. miRNA expression was normalized (counts per millions), with hierarchical clustering computed using Euclidean distance method. (D) Pathway analysis of differentially regulated miRNA gene targets in days 4 versus 7 is shown. miRNA gene targets for the top 31 differentially expressed miRNAs were determined using miRTarBase. String-db was used to examine fit of miRNA gene targets to GO biological process pathways. Various immune system processes (blue bars) and apoptotic pathways (red bars) were enriched (FDR < 0.05). (E) Venn diagrams of enriched apoptosis and immune-specific GO biological processes for day 4 and day 7 upregulated miRNA gene targets is shown. Day 4 showed upregulation miRNAs targeting genes involved in immune response pathways, and day 7 showed miRNAs targeting genes associated with apoptosis pathways. (F) Bar plot of genes targeted by multiple differentially expressed miRNAs. Gene targets of miRNAs upregulated in day 4 and day 7 are shaded in blue and orange, respectively. Only gene targets with >1 miRNA hit are shown. Abbreviations: BP, biological process; fc, fold change; FDR, false discovery rate; GO, gene ontology; miRNA, micro RNA; sEV, small extracellular vesicle.

3.5. T cell sEV miRNAs may serve as candidate biomarkers of ACR

To understand whether the NGS data analysis may help identify novel candidate biomarkers of ACR, we developed stem loop RT-qPCR assays for 3 candidate miRNAs that were detected by principal component analysis–miR-let7i, miR-21a, miR-101b. We did syngeneic (n = 5) and full major histocompatibility complex (MHC) mismatch (BALB/c into C57BL/6) (n = 5) transplants, harvested T cell sEV RNA cargoes, and performed RT-qPCR for the 3 candidate miRNAs. Expression patterns in the Rejection (n = 5) versus Control (n = 5) arms showed significant upregulation of miR-let7i and miR-21a (Fig. 5A, B), with similar trend for miR-101b (Fig. 5C).

Figure 5.

T cell sEV stem loop RT-qPCR assays for 3 candidate miRNA biomarkers determined by NGS differential expression analysis is shown. Animals from the Rejection (n = 5) and Control (n = 5) arms were sacrificed on POD 7, and RNA from CD3 sEVs was extracted and analyzed for miRNA expression levels by stem loop RT-qPCR for: (A) miR-let7i, (B) miR-21a, and (C) miR-101. Average relative fold expression with 95% confidence interval in the Control arm compared to the Rejection arm is shown. By t test, miR-let7i (P = .03) and miR-21a (P = .046) expression was significantly higher in T cell sEVs with allograft rejection versus syngeneic controls. miR-101b showed a higher level of expression but was not significant (P = .23). Abbreviations: miRNA, micro RNA; NGS, next generation sequencing; POD, postoperative day; RT-PCR, reverse transcription polymerase chain reaction; RT-qPCR, reverse transcription-quantitative polymerase chain reaction; sEV, small extracellular vesicle.

3.6. T cell sEVs released secondary to alloactivation independently mediate donor tissue-specific apoptosis

Given the findings suggesting upregulation of alloreactivity and apoptosis pathways in T cell sEVs with ACR, we tested whether T cell sEVs can independently mediate allogeneic donor-specific cytotoxicity. BALB/c (donor) or C57BL/6 (recipient) cardiomyocytes were incubated with either total plasma sEVs, T cell sEVs, CD3 antibody unbound sEVs, or plasma secretome in sEV preparations of day 7 samples from Rejection or Control arms (Fig. 6A). Cytotoxicity was assessed by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay. Control arm sEV samples and secretome did not mediate apoptosis of either BALB/c (allogeneic) or C57BL/6 (syngeneic) cells (Fig. 6B, E) compared to DNase I-treated positive controls (Fig. 6D, G). C57BL/6 cells treated with sEV fractions and secretome from the Rejection arm also did not show any apoptosis (Fig. 6F). However, BALB/c cells treated with Rejection arm total plasma sEVs and CD3⁺ sEVs showed positivity by TUNEL assay (Fig. 6C). Apoptosis rates were significantly higher for BALB/c cells treated with Rejection arm CD3⁺ sEVs compared to any other study group (P <.0001, Fig. 6H) except for Rejection arm total plasma sEVs (P = .28). Cytotoxic T cell–mediated apoptosis occurs via a granule (containing granzymes and perforin)-dependent exocytosis pathway and the Fas-FasL pathway. Therefore, we assessed if T cell sEVs express these markers. On western blot, expression of FasL, granzyme B, and perforin were seen in the enriched plasma T cell sEV subpopulations in the naive, syngeneic recipient, and allogeneic recipient animals, but higher expression levels were seen in the Rejection arm (Fig. 6I, J). CD38 and MHC II, markers of T cell activation status, were also upregulated in the Rejection arm T cell sEV subset. Collectively, this suggests that a subset of circulating T cell sEVs during ACR carry functional capabilities to mediate donor-specific cytotoxicity in vitro and express proapoptotic and activation markers like their cytotoxic T cell counterparts.

Figure 6. Circulating T cell sEVs enriched from Rejection arm animals independently mediate donor cardiomyocyte specific cytotoxicity.

(A) Schematic of plasma components analyzed for potential to mediate donor cardiomyocyte cytotoxicity is shown. Total plasma sEVs from POD 7 in Rejection and Control arms was harvested. From this, T cell sEVs were enriched using anti-CD3 antibody beads. The unbound sEVs with the plasma secretome was then ultracentrifuged to obtain the secretome and unbound sEV fraction. For apoptosis assay, donor BALB/c cardiomyocytes or recipient C57BL/6 cardiomyocytes were incubated with the following isolated fractions from either the Rejection arm or Control arm: total plasma sEVs (1×10¹⁰ sEVs), CD3 sEVs (bound fraction) (1.5×10⁸ sEVs), bead unbound sEVs, and remnant secretome. (B) Confocal microscopy images of TUNEL assay for BALB/c cells incubated with Control arm total plasma sEVs, CD3 sEVs, unbound sEVs, and secretome failed to show apoptosis. (C) BALB/c cells incubated with Rejection arm total plasma sEVs and CD3 sEVs showed >45% apoptosis, but minimal (<6%) apoptosis was seen with unbound sEVs and secretome. (D) BALB/c cells treated with DNase I (positive control) showed near 100% apoptosis. (E) C57BL/6 cells incubated with Control arm plasma components showed <5% apoptosis. (F) C57BL/6 cells incubated with Rejection arm plasma sEVs, CD3 sEVs, unbound sEVs, and secretome failed to show any apoptosis. (G) C57BL/6 cells treated with DNase I showed >97% apoptosis (positive control). (H) Summary of apoptosis assay is shown. Results of 2 independent experiments with percentage of apoptosis positive cells is displayed. By Fisher exact test, apoptosis rates were significantly higher in BALB/c cells treated with Rejection arm CD3 sEVs compared to all other tested conditions (p<0.0001), except similar apoptosis rates seen between Rejection arm total plasma sEVs and CD3 sEVs (p=0.28). (I) CD3 sEVs from C57BL/6 wild type, BALB/c wild type, Control arm day 7, and Rejection arm day 7 were assessed for expression of pro-apoptotic proteins known to be critical for T cell mediated cytotoxicity. All study groups showed expression of Fas ligand (FasL), perforin-1, granzyme B, TCR, CD38, MHC II, but higher levels of these pro-apoptotic markers and activation markers were seen in the Rejection arm. Flotillin-1 and TSG101 are exosome markers. One of two experiments is shown. (J) Relative expression of pro-apoptotic markers and activation markers normalized to expression of exosome marker TSG101 is shown. Rejection arm recipients showed higher expression of these markers in their circulating T cell sEVs.

3.7. Circulating T cell sEVs can be detected in clinical heart transplantation

To understand whether similar changes may occur in circulating T cell sEV physiology during ACR in the clinical setting, we investigated this platform in heart transplant patients. In 3 patients with ACR episodes diagnosed by EMB, time-matched blood samples were obtained for T cell sEV profiling (Supplementary Table S1). Grade ≥2 ACR is clinically significant, mandating treatment with additional immunosuppression and repeating EMB to confirm reversal of rejection to grade 0/1. Plasma sEVs were characterized by electron microscopy (Fig. 7A), NTA (Fig. 7B, Supplementary Files 7–9), and western blot for EV markers (Fig. 7C) per MISEV guidelines.¹⁷ Methodology for enrichment of CD3-expressing sEVs in the mouse model was translated to the clinical setting. Anti-CD3 antibody bead-bound sEVs were eluted and analyzed by NTA to confirm enrichment of intact sEVs (Fig. 7D, Supplementary File 10). This fraction was studied by western blot for confirmation of expression of T cell markers, with their absence in the unbound fraction (Fig. 7E, F).

Figure 7.

Circulating T cell sEVs can be enriched in heart transplant patients. sEVs were isolated from heart transplant patient (n = 3) peripheral blood samples (n = 14) at different time points posttransplant. Sample collection was time-matched to endomyocardial biopsy, the results of which are shown next to the POD time point. (A) Electron microscopy of plasma sEVs in Subject 2 is shown. (B) NTA light scatter analysis of sEVs isolated in Subjects 1 to 3 is shown with mean particle concentration and size. (C) sEV protein markers were assessed by western blot for expression of exosome markers TSG101, flotillin-1, CD63, and Alix-1 and for absence of cytochrome c, calnexin-1, and ApoE, per MISEV guidelines. Representative western blot for Subject 2 is shown. (D) CD3 sEVs were enriched from plasma sEVs using anti-CD3 antibody conjugated beads, and the bead-bound fraction was eluted and analyzed on NTA to confirm enrichment of intact sEVs. Representative NTA for 3 time points (PODs 1, 13, and 98) in Subject 2 is shown with particle number and size distribution. Given we started with 1×10¹⁰ sEVs for bead isolation and eluted between 0.9 to 1.4×10⁸ sEVs as enriched T cell sEV fraction, it represents potential enrichment of T cell sEV signal by ~ 60- to 100-fold compared to whole plasma sEV analysis. (E) Western blot analysis of CD3 antibody bead-bound sEVs was performed to validate expression of T cell markers. Representative analysis for Subject 2 at PODs 1, 13, and 98 are shown along with Jurkat cell positive control for expression of CD8 in the T cell sEV fraction. Flotillin-1, an exosome marker, expression is also shown. (F) Western blot analysis of CD3 antibody unbound sEVs was performed to validate absence of T cell markers. Representative analysis for Subject 2 at 3 postoperative time points is shown for expression of CD3 and flotillin-1. CD3 expression was not seen in the unbound sEV fraction, but flotillin-1 (exosome marker) expression was detected. Jurkat EVs and Jurkat cell positive controls are shown. (G-I) Western blot analysis of T cell sEVs in Subjects 1 (G), 2 (H), and 3 (I) for expression of T cell inflammatory and proapoptotic proteins is shown, including perforin-1, granzyme B, IFNγ, FasL, CD38, and MHC II (HLA-DR). EMB histology data (ACR 0 versus 1 versus 2) is also shown. Grade 2 ACR by EMB correlated with increased expression of these T cell activation marker proteins. Flotillin-1 is an exosome marker protein. (J-L) Stem loop RT-qPCR for miR-let7i, miR-21a, and miR-101b expression in T cell sEVs in Subjects 1 (J), 2 (K), and 3 (L) at EMB-matched postoperative time points is shown. Relative expression of candidate miRNAs compared to POD 1 baseline (set to 1) is shown. Correlation of miRNA expression with EMB histology data showed increased expression of these miRNAs in T cell sEV fraction at time points of grade 2 ACR. Abbreviations: ACR, acute cellular rejection; ApoE, apolipoprotein E; EMB, endomyocardial biopsy; FasL, Fas ligand; HLA-DR, human leukocyte antigen-DR isotype; IFN, interferon; MHC, major histocompatibility complex; MISEV, Minimal Information for Studies of Extracellular Vesicles; NTA, nanoparticle tracking analysis; POD, postoperative day; RT-qPCR, reverse transcription-quantitative polymerase chain reaction.

In all 3 patients, T cell intra-sEV cargoes showed upregulation of IFNγ, granzyme B, perforin, FasL, human leukocyte antigen class II, and CD38 by western blot at time points of grade 2 ACR (Fig. 7G–I). Resolution of grade 2 ACR led to corresponding decrease in expression of these activation markers in the T cell sEV subset. Next, we assessed whether the 3 tested candidate T cell sEV miRNAs in the rodent ACR model may also be differentially regulated in clinical heart transplantation. By stem loop RT-qPCR, miR-let7i, miR-101b, and miR-21a expression in circulating T cell EVs demonstrated upregulation of the candidate miRNAs with grade 2 ACR episodes in all 3 patients (Fig. 7J–L, Supplementary Table S2). Grade 0 ACR by EMB correlated with the baseline signal assessed on POD 1. Taken together, these findings suggest that the circulating T cell sEV physiology of ACR detailed in the rodent model may also occur during ACR in the clinical setting.

Given the detection of allogeneic cytotoxicity of T cell sEVs, and that they express TCR mRNA and protein, it raises the possibility that targeted donor specificity may be achieved via functional T cell receptors on CD3 sEVs. Therefore, we assessed whether T cell clones with intact complementarity determining region 3 sequences can be sequenced from the circulating T cell sEV subset. In 4 heart transplant patients during time points of grade 0 or grade 2 ACR, productive TCR sequences were successfully identified, with 38 unique functional clones (Table 3). Clustering analysis of these 38 sequences against a large reference data set containing 20 million TCRs demonstrated that only 1 TCR sequence clustered, indicating that the other 37 TCRs were unique to this cohort (Table 4).²² This suggests that T cell sEVs may carry clone-specific TCRs.

Table 3.

Past-filter TCR sequences with intact complementarity determining region 3 in 4 heart transplant patients with moderate grade 2 ACR events is shown for 11 studied time points. TCR sequences with nonproductive CDR3, failed TRBV calls or missing conserved motifs were removed.

CDR3 sequence	TRB variable-gene	Clonal frequency	Rank	Sample name
CASSQSRPIIHHNSPLHF	TRBV3-1*01	0.222222222222222	1	Patient 2_POD-0_NO_REJ_TCRB.tsv
CASRSILQHARNTIYF	TRBV21-1*01	0.0769230769230769	1	Patient 2_POD13_MOD_REJ_TCRB.tsv
CASSLCHPILHHQPQHF	TRBV3-1*01	0.125	1	Patient 2_POD_98_NO_REJ_TCRB.tsv
CASSYLYPISETQYF	TRBV3-1*01	0.0625	0.5	Patient 2_POD_98_NO_REJ_TCRB.tsv
CTREQSRANVLTF	TRBVA-1*01	0.0625	0.5	Patient 2_POD_98_NO_REJ_TCRB.tsv
CASSRFGPVMQDTQYF	TRBV3-1*01	0.0714285714285714	1	Patient 4_POD30_MOD_ACR_TCRB.tsv
CASSRFGPVMQDTQYF	TRBV3-1*01	0.0357142857142857	0.5	Patient 4_POD_ 30MOD_ACR_TCRB.tsv
CASRDRREQFF	TRBV21-1*01	0.0357142857142857	0.5	Patient 4_POD30_MOD_ACR_TCRB.tsv
CASSPRAPHSDNEQFF	TRBV3-1*01	0.0357142857142857	0.5	Patient 4_POD30_MOD_ACR_TCRB.tsv
CASSEDLGGKTYGYTF	TRBV6-1*01	0.0357142857142857	0.5	Patient 4_POD30_MOD_ACR_TCRB.tsv
CASSCITKCRTYGYTF	TRBV7-7*01	0.0357142857142857	0.5	Patient 4_POD30_MOD_ACR_TCRB.tsv
CASSSPEELESRANVLTF	TRBV25-1*01	0.0357142857142857	0.5	Patient 4_POD30_MOD_ACR_TCRB.tsv
CAQQRHHSWHHQPQHF	TRBV3-1*01	0.125	1	Patient 4_POD65_NO_ACR_1_TCRB.tsv
CASTLIVSHKVGHVLTF	TRBV5-5*01	0.0625	0.5	Patient 4_POD65_NO_ACR_1_TCRB.tsv
CLTECICFLCLTYRVVHF	TRBV15*01	0.0625	0.5	Patient 4_POD65_NO_ACR_1_TCRB.tsv
CASSLRAKCQSRANVLT	TRBV7-7*01	0.125	0.9	Patient 4_POD65_NO_ACR_2_TCRB.tsv
CASSSTYPIFGHQPQHF	TRBV3-1*01	0.125	0.9	Patient 4_POD65_NO_ACR_2_TCRB.tsv
CASSEIGYHDKVGHVLTF	TRBV7-9*01	0.0625	0.4	Patient 4_POD65_NO_ACR_2_TCRB.tsv
CASSTQVSGLTGELFF	TRBV3-1*01	0.0625	0.4	Patient 4_POD65_NO_ACR_2_TCRB.tsv
CASSGKGPILHHQPQHF	TRBV3-1*01	0.0625	0.4	Patient 4_POD65_NO_ACR_2_TCRB.tsv
CASSQRGPIHNSPLHF	TRBV3-1*01	0.0833333333333333	0.625	Patient 3_POD-0_NO_REJ_TCRB.tsv
CASSLAAPIIHHNSPLHF	TRBV3-1*01	0.0833333333333333	0.625	Patient 3_POD-0_NO_REJ_TCRB.tsv
CASSYNLATVHNEKLFF	TRBV21-1*01	0.0833333333333333	0.625	Patient 3_POD-0_NO_REJ_TCRB.tsv
CASSVWHHQPQHF	TRBV3-1*01	0.0833333333333333	0.625	Patient 3_POD-0_NO_REJ_TCRB.tsv
CASSQFFIHHNSPLHF	TRBV3-1*01	0.166666666666667	0.666	Patient 3_POD 16_MOD_REJ_TCRB.tsv
CSVSKKLAYSRANVLTF	TRBV29-1*01	0.166666666666667	0.666	Patient 3_POD-16_MOD_REJ_TCRB.tsv
CASVKRVRKTYGYTF	TRBV5-5*01	0.166666666666667	0.666	Patient 3_POD-16_MOD_REJ_TCRB.tsv
CASSVKYPIHNSPLHF	TRBV3-1*01	0.125	0.75	Patient 3_POD-74_NO_REJ_TCRB.tsv
CASSSLSPIFGHQPQHF	TRBV3-1*01	0.125	0.75	Patient 3_POD-74_NO_REJ_TCRB.tsv
CASSELCPICKKNIQYF	TRBV3-1*01	0.0909090909090909	1	Patient 3_POD9_MOD_REJ_TCRB.tsv
CATSMAKRCDKVGNVLTF	TRBV15*01	0.0526315789473684	0.583	Patient 1_POD38_MOD_ACR_TCRB.tsv
CASSCRVKGTKSGQRPDF	TRBV4-2*01	0.0526315789473684	0.58	Patient 1_POD38_MOD_ACR_TCRB.tsv
CSVLHHIQSRANVLTF	TRBV29-1*01	0.0526315789473684	0.58	Patient 1_POD38_MOD_ACR_TCRB.tsv
CASRRPLATRNEKLFF	TRBV21-1*01	0.0526315789473684	0.583	Patient 1_POD38_MOD_ACR_TCRB.tsv
CASKMYWDTQSRANVLT	TRBV5-2*01	0.0526315789473684	0.583	Patient 1_POD38_MOD_ACR_TCRB.tsv
CASSRIYPIYHADTQYF	TRBV3-1*01	0.0526315789473684	0.583	Patient 1_POD38_MOD_ACR_TCRB.tsv
CASSPPNPVMQDTQYF	TRBV3-1*01	0.0833333333333333	0.75	Patient 1_POD65_NO_ACR_1_TCRB.tsv
CASSPTTPILHHQPQHF	TRBV3-1*01	0.0833333333333333	0.75	Patient 1_POD65_NO_ACR_1_TCRB.tsv

Table 4.

GIANA query of T cell EV-derived TCRs against a large reference human TCR database.

CDR3	Cluster ID	TRBV gene	Clonal frequency	Sample ID	Query status
CASRDRREQFF	2	TRBV21-1*01	0.0001245	COVID19:BS-EQ-0006-T2-replacement_TCRB.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	4.77E-05	COVID19:BS-EQ-0009-T2-replacement_TCRB.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	3.77E-05	COVID19:BS-EQ-44-T0_BS-GIGI_11-replacement_TCRB.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	2.32E-05	COVID19:BS-GIGI_46-replacement_TCRB.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	5.95E-05	COVID19:BS-HS-18_TCRB.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	1.72E-05	COVID19:KHBR20-00193_TCRB.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	3.74E-05	LungCancer_MDanderson2019:MDA-1636-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	1.92E-05	LungCancer_MDanderson2019:MDA-1908-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	1.89E-05	LungCancer_MDanderson2019:MDA-1963-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	3.79E-05	LungCancer_MDanderson2019:MDA-2025-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	3.49E-05	LungCancer_MDanderson2019:MDA-2175-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	2.92E-05	LungCancer_MDanderson2019:MDA-2244-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	2.19E-05	LungCancer_MDanderson2019:MDA-2325-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	2.62E-05	LungCancer_MDanderson2019:MDA-3203-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	1.48E-05	LungCancer_MDanderson2019:MDA-3212-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	0.00012317	LungCancer_MDanderson2019:MDA-4378-C.tsv	ref
CASRDRREQFF	2	TRBV21-1*01	0.03571429	JAT_MOD_ACR_TCRB.tsv	query

4. Discussion

There remains a critical need for development of time-sensitive, accurate biomarkers for ACR. In a recent meta-analysis reviewing literature on efficacy of T cell–mediated rejection therapy in kidney transplant patients, the incidence of index subclinical acute rejection by biopsy was 30%, and clinical biopsy-proven acute rejection was noted in 16% of cases.²³ Importantly, follow-up biopsy showed pooled proportion of persistent ≥ Banff borderline acute rejection of 0.39 (95% confidence interval, 0.26–0.53).²³ Other studies also attest to the importance of constant surveillance for ACR.^24,25 In heart transplantation, ACR still remains a major cause of morbidity and mortality, especially during the first year posttransplant.²⁶ EMB remains the mainstay and gold standard for ACR diagnosis and surveillance, although biomarker platforms such as peripheral blood mononuclear cell gene expression profiling²⁷ and cell free DNA assays have made inroads.^28–30 These platforms, however, have not replaced EMB and have limitations, indicating the need for development of other noninvasive platforms. EMB itself also has some limitations—biopsy grading can be subjective and prone to interoperator differences;³¹ it is not practical for frequent repeated testing like blood sampling, is invasive, resource intensive, costly, and associated with risks.³²

Previously, we investigated the diagnostic utility of circulating donor tissue-specific sEV profiles as a noninvasive diagnostic of ACR in models of islet cell transplantation, heart transplantation, and lung transplantation.^13–15 This showed that donor sEV profiles from day 2 posttransplant predicted ACR with 100% accuracy. In this context, we hypothesized that alloactivation of T cells during ACR would also result in a time-specific T cell sEV “footprint” in peripheral blood, with their intra-sEV cargoes contributing to ACR pathophysiology. The findings of this study support this hypothesis, raising the possibility that a composite assay characterizing specific intra-sEV cargoes of donor tissue sEVs and T cell sEVs may enable development of a novel biomarker platform for ACR surveillance. Interestingly, we noted that though peripheral blood T cell counts are not significantly altered with ACR, there is a distinct ACR-associated surge in circulating T cell sEVs, suggesting that sEV profiles may reflect functional status of their cellular counterparts rather than the viable or circulating cellular mass. Accordingly, resolution of grade 2 ACR in the 3 patients led to reversal of T cell sEV markers to baseline levels. This suggests that the T cell sEV platform may also enable noninvasive monitoring of efficacy of treatment of clinically grade ≥2 ACR episodes, which currently mandates repeat EMB to confirm successful treatment of ACR. In future, this idea will need to be studied in the clinical setting.

In addition to changes in quantitative profiles, we hypothesized that ACR would also lead to changes in T cell sEV cargoes. In support of this, we noted increased expression of proinflammatory T cell markers IL2, TCR, CD38, MHC II, and IFNγ in T cell sEVs in the mouse and in human heart transplant ACR models. Furthermore, NGS for small RNAs revealed upregulation of specific miRNAs in T cell sEVs. For 3 candidate miRNA biomarkers, miR-let7i, miR-21a, and miR-101b, stem loop RT-qPCR assays showed upregulation of these markers during grade 2 ACR episodes (Fig. 7J–L). miR-let7i expression in T cells has been shown to promote helper T cell phenotype (IFNγ production),³³ and in circulating exosomes, miR-let7i has been implicated in inhibition of regulatory T cells.^34,35 Increased expression of miR-21 and miR-101 with ACR has been reported.^36,37 Collectively, these findings support our concept that changes in protein and RNA cargoes of T cell sEVs may serve as candidate biomarkers of ACR.

We believe that the described T cell sEV platform also has physiological and mechanistic implications. Several results of our investigation suggest that T cell sEVs play a role in the pathophysiology associated with ACR: (1) time-specific changes in circulating T cell sEV profiles were seen in the Rejection arm only; (2) ACR led to increased expression of alloactivation markers inside T cell sEVs; (3) pathway analysis of T cell sEV miRNAs showed sequential upregulation of immune regulatory pathways and apoptotic pathways on days 4 and 7 posttransplant; (4) miR target analysis showed regulation of genes involved in apoptosis pathways, including HDAC4,³⁸ PTEN,³⁹ BCL2, MEF2c,^38,40 SMAD7,⁴¹ and CAB39;⁴² (5) T cell sEVs expressed proteins important in mediating targeted apoptosis by cytotoxic T cells, including TCR, FasL, granzyme B, and perforin; (6) only T cell sEVs from the Rejection arm mediated apoptosis of donor-specific BALB/c cardiomyocytes on TUNEL assay; and (7) intact TCR complementarity determining region 3-unique clones were sequenced from T cell sEVs. Taken together, these findings raise the possibility that in addition to cell-cell contact-mediated apoptosis of donor tissue, alloreactive T cell clones may potentiate donor tissue inhibition/injury and cell death even before they migrate to the allograft via their sEVs.

Although outside the scope of this manuscript, we believe that the target specificity of T cell sEVs is mediated via the expressed TCR, like their cellular counterparts.^43–47 If so, this opens the window for studying specific T cell clonal subpopulations activated during ACR. Although such studies have been performed directly from allograft biopsies or from recipient peripheral blood mononuclear cell samples after in vitro stimulation using irradiated donor cells,^46,47 to our knowledge, they have not been studied in sEVs. Having detected both TCR protein and mRNA in T cell sEVs, we were able to perform sequencing of intact TCR clones utilizing the sEV mRNA cargo. Future studies on T cell sEVs may enable characterization of patient-specific T cell receptor clones that may play dominant roles during ACR episodes.

In summary, we detail a novel noninvasive cell-specific sEV biomarker platform with diagnostic potential for monitoring ACR. Circulating T cell sEV profiles are significantly altered in a time-specific manner during ACR, attesting to their potential in transplant diagnostics. Furthermore, pathway and cargo analysis of T cell sEVs and their ability to independently mediate apoptosis of donor cardiomyocytes suggest a functional role for T cell sEVs in the ACR process. Future studies investigating mechanistic roles of specific mRNA and miRNA networks in T cell sEVs may provide novel therapeutic windows for treating ACR.

5. Study limitations

T cell sEV NGS was performed to understand miRNA cargo changes comparing day 4 versus day 7 posttransplant samples in the Rejection arm, but NGS was not performed for the Control arm (syngeneic transplants). The results of the apoptosis assay will need in vivo investigation in the future to validate any mechanistic implications. Although T cell sEV cargoes were altered with ACR in clinical setting, a study with larger sample size and longer follow-up will be needed to understand the biomarker potential of T cell sEVs in heart transplantation.

Supplementary Material

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.ajt.2023.10.021.

Abbreviations

ACR

acute cellular rejection

ANOVA

analysis of variance

EMB

endomyocardial biopsy

EV

extracellular vesicle

FACS

fluorescence activated cell sorting

fc

fold change

MHC

major histocompatibility complex

MISEV

Minimal Information for Studies of Extracellular Vesicles

miRNA

microribonucleic acid

NGS

next generation sequencing

NTA

nanoparticle tracking analysis

POD

postoperative day

RT-PCR

reverse transcription polymerase chain reaction

RT-qPCR

reverse transcription-quantitative polymerase chain reaction

sEV

small extracellular vesicle

TCR

T cell receptor

TUNEL

terminal deoxynucleotidyl transferase dUTP nick end labeling

Data availability statement

For request to access to further data, please contact the senior author.