Double-negative T cells have a reparative role after experimental severe ischemic acute kidney injury

Abstract

T cells mediate organ injury and repair. A proportion of unconventional kidney T cells called double-negative (DN) T cells (TCR⁺ CD4⁻ CD8⁻), with anti-inflammatory properties, were previously demonstrated to protect from early injury in moderate experimental acute kidney injury (AKI). However, their role in repair after AKI has not been studied. We hypothesized that DN T cells mediate repair after severe AKI. C57B6 mice underwent severe (40 min) unilateral ischemia-reperfusion injury (IRI). Kidney DN T cells were studied by flow cytometry and compared with gold-standard anti-inflammatory CD4⁺ regulatory T cells (Tregs). In vitro effects of DN T cells and Tregs on renal tubular epithelial cell (RTEC) repair after injury were quantified with live-cell analysis. DN T cells, Tregs, CD4, or vehicle were adoptively transferred after severe AKI. Glomerular filtration rate (GFR) was measured using fluorescein isothiocyanate (FITC)-sinistrin. Fibrosis was assessed with Masson’s trichrome staining. Profibrotic genes were measured with qRT-PCR. Percentages and the numbers of DN T cells substantially decreased during repair phase after severe AKI, as well as their activation and proliferation. Both DN T cells and Tregs accelerated RTEC cell repair in vitro. Post-AKI transfer of DN T cells reduced kidney fibrosis and improved GFR, as did Treg transfer. DN T cell transfer lowered transforming growth factor (TGF)β1 and α-smooth muscle actin (αSMA) expression. DN T cells reduced effector-memory CD4⁺ T cells and IL-17 expression. DN T cells undergo quantitative and phenotypical changes after severe AKI, accelerate RTEC repair in vitro as well as improve GFR and renal fibrosis in vivo. DN T cells have potential as immunotherapy to accelerate repair after AKI.

NEW & NOTEWORTHY Double-negative (DN) T cells (CD4⁻ CD8⁻) are unconventional kidney T cells with regulatory abilities. Their role in repair from acute kidney injury (AKI) is unknown. Kidney DN T cell population decreased during repair after ischemic AKI, in contrast to regulatory T cells (Tregs) which increased. DN T cell administration accelerated tubular repair in vitro, while after severe in vivo ischemic injury reduced kidney fibrosis and increased glomerular filtration rate (GFR). DN T cell infusion is a potential therapeutic agent to improve outcome from severe AKI.

INTRODUCTION

Acute kidney injury (AKI) is a common and serious clinical problem resulting in high morbidity and mortality worldwide (1). Impaired recovery from AKI can lead to kidney fibrosis and transition to chronic kidney disease (CKD) (2). Although prior studies mostly focused on the prevention of early injury (3), understanding molecular and cellular mechanisms of repair and recovery after AKI is clinically important given that most patients are diagnosed after AKI has occurred (4).

Among the many cellular and molecular pathways involved in the AKI repair process (3, 5–9), immune responses mediated by T cells have been proposed as one of the important pathways (10, 11). Long-term increase in numbers, immunophenotypical changes, and transcriptomic reprogramming of T cells have been demonstrated in previous studies (12–14), highlighting their potential role in AKI repair or CKD transition. Furthermore, a minor proportion of kidney CD4⁺ T cells, regulatory T cells (Tregs) that have anti-inflammatory properties not only have a protective role in early injury (15), but also have a reparative role in AKI to CKD transition (16).

An unconventional T cell subset, double-negative (DN) T cells that do not express either CD4 or CD8 exist in kidneys (17–20). Although they are rarely present in lymphoid tissue and peripheral blood, there are significant proportions of DN T cells among total αβ T cells in steady state as well as post-AKI kidneys (19). DN T cells exhibited a protective role with an anti-inflammatory property in prevention from moderate early injury (19), however little is known about their role in repair after AKI, particularly after clinically significant severe injury. We therefore hypothesized that kidney DN T cells change after severe AKI and can directly participate in repair. We studied the effects of DN T cells after severe AKI, effects on renal epithelial cells in vitro as well as in vivo AKI, comparing them to the “gold standard” anti-inflammatory CD4⁺ Tregs (18, 20).

MATERIALS AND METHODS

Mice

Seven-week-old male C57BL/6J wild-type (WT) mice were purchased from Jackson Laboratory (Bar Harbor, ME) and housed under specific pathogen-free conditions at the Johns Hopkins University animal facility. Eight to nine week old mice were used for experiments using WT mice. Twelve-week-old Fasl^gld/J male mice were used as donors for DN T cell isolation as previously described (19, 21). All experiments were performed using experimental protocols approved by the Animal Care and Use Committee of Johns Hopkins University and reported in compliance with the ARRIVE guideline (22).

Severe Ischemic AKI Model

WT mice were anesthetized with pentobarbital (75 mg/kg; Akorn, Lake Forest, IL) injection intraperitoneally. Mice were placed onto a thermostatically controlled heating table after shaving of abdominal hair. Abdominal midline incision was performed, and left renal pedicles were dissected and clamped for 40 min using a microvascular clamp (Roboz Surgical Instrument, Gaithersburg, MD) to induce severe ischemia. The clamps were released from renal pedicles after 40 min, and the left kidneys were visually inspected to confirm reperfusion. Mice were kept well hydrated with 1 mL of warm sterile 0.9% saline and at a constant body temperature (37°C) during the surgery. After being sutured, mice were allowed to recover with free access to chow and water.

Assessment of Kidney Function

Since serum creatinine is a less sensitive measure of glomerular filtration rate (GFR) in the unilateral ischemia-reperfusion injury (IRI) model due to the remaining functional contralateral kidney (23), we directly measured glomerular filtrate rate (GFR) to measure kidney function. GFR was measured by transcutaneous fluorescein isothiocyanate (FITC)-sinistrin (inulin analog) with a fluorometer device (MediBeacon, St. Louis, MO) at baseline, 24 h, 1 wk, 2 wk, and 3 wk after reperfusion (24). Briefly, mice were anesthetized with isoflurane (Piramal, Maharashtra, India) and oxygen under an isoflurane vaporizer system (VetFlo, Kent Scientific, Torrington, CT). The background fluorescence signal of skin was recorded for 5 min, and subsequently 0.07 mg/g body wt of FITC-sinistrin (MediBeacon) was injected retro-orbitally. Mice were immediately transferred to separate cages to record FITC-sinistrin clearance in dark. After 1.5 h, the devices were gently detached from conscious mice, and raw data from the devices were collected using MB Lab Software (MediBeacon). GFR was calculated using a previously established three-compartment model (25) by Studio2 Software (MediBeacon).

Tissue Histological Analysis

At 3 wk after the surgery, mice were anesthetized with intraperitoneal injection of ketamine (130 mg/kg; VetOne, Boise, ID) and xylazine (7 mg/kg; Akorn) mixture. Mice were exsanguinated, and kidneys were collected. Left kidney tissues were fixed with 10% buffered formalin followed by paraffin embedding. Kidney sections were subsequently stained with Masson’s trichrome staining. A renal pathologist, blinded to the study groups, scored the degree of fibrosis from the kidney sections.

Isolation of Kidney Mononuclear Cells

For kidney mononuclear cell isolation, postischemic kidneys and contralateral kidneys were collected at 1 wk and 3 wk after the IRI surgery. Uninjured intact kidneys from naïve mice were also collected for steady-state controls. Kidney mononuclear cells (KMNCs) were isolated using Percoll density gradient protocol described previously (26). Briefly, decapsulated kidneys were incubated in 2 mg/mL collagenase D (Roche, Basel, Switzerland) solution for 30 min at 37°C. Samples were strained through 70-μm cell strainer (BD Biosciences, Franklin Lakes, NJ), washed, and resuspended in 40% Percoll (GE Healthcare, Chicago, IL) followed by gentle overlaying onto 80% Percoll. After centrifugation at 1,800 g for 30 min in brake-off mode at room temperature, KMNCs were collected from the interface between 40% and 80% Percoll. Collected cells were washed and resuspended with Roswell Park Memorial Institute (RPMI) 1640 media (Thermo Fisher Scientific, Waltham, MA) containing 5% fetal bovine serum (FBS, Thermo Fisher Scientific). Cells were counted on a hemocytometer using trypan blue (Thermo Fisher Scientific) under a microscope (IMT-2, Olympus, Tokyo, Japan).

Spectral Flow Cytometry

Cells were washed once with phosphate-buffered saline (PBS) and stained with viability dye Zombie NIR Fixable Viability (BioLegend, San Diego, CA) for 15 min at room temperature. After washing with Cell Staining Buffer (BioLegend), cells were preincubated with anti-CD16/CD32 Fc receptor blocking antibody (S17011E, BioLegend) for 15 min to prevent nonspecific antibody binding. Subsequently, surface staining was performed with surface staining antibody cocktail in 50 μL of BD horizon Brilliant Stain buffer for 30 min at 4°C: Pacific blue anti-CD44 (IM7, BioLegend), BV510 anti-CD8 (53-6.7, BioLegend), BV570 anti-CD45 (30-F11, BioLegend), BV605 anti-CD69 (H1.2F3, BioLegend), BV650 anti-NK1.1 (PK136, BioLegend), BV711 anti-PD1 (29 F.1A12, BioLegend), BV785 anti-TCRβ (H57-597, BioLegend), Alexa Fluor 532 anti-CD3 (17A2, Thermo Fisher Scientific), PE/Dazzel 594 anti-T cell immunoreceptor with Ig and ITIM domains (TIGIT) (1G9, BioLgend), PE-Cy5 anti-CD122, PE-Cy5.5 anti-CD25 (PC61.5, Thermo Fisher Scientific), PE-Cy7 anti-Ly49 (14B11, BioLgend), Alexa Fluor 647 anti-TCRγδ (GL3, BioLegend), APC-R700 anti-CD62L (MEL-14, BD Biosciences), and APC-Fire810 anti-CD4 (GK1.5, BioLegend). Cells were fixed and permeabilized with Foxp3/Transcription Factor Staining kit (Thermo Fisher Scientific) for 30 min at room temperature and washed with permeabilization/wash buffer (Thermo Fisher Scientific). Intracellular staining was conducted in 50 μL of permeabilization/wash buffer with intracellular staining antibody cocktail for 30 min at room temperature: BV421 anti-Ki67 (16A8, BioLegend), PerCP-efluor 710 anti-FoxP3 (FJK-16S, Thermo Fisher Scientific).

T Cell Activation and Intracellular Cytokine Analysis

To measure intracellular cytokines, KMNCs were stimulated with premixed leukocyte activation cocktail (BioLegend) containing phorbol 12-myristate-13-acetate, ionomycin, and brefeldin A. After surface staining followed by permeabilization and fixation as described earlier, cells were stained with the following intracellular antibodies, BV421 anti-Ki67 (16A8, BioLegend), Alexa Fluor 488 anti-TNFα (MP6-XT22, BioLegend), Alexa Fluor 532 anti-IL-2 (JES6-5H4, BD Biosciences), PerCP-efluor 710 anti-FoxP3 (FJK-16S, Thermo Fisher Scientific), PE anti-IL-10 (JES5-16E3, Biolegend), PE-Cy5 anti-INFγ (XMG1.2, Abcam), PE-Cy7 anti-IL17A (TC11-18H10, Biolegend), Alexa Fluor 647 anti-transforming growth factor (TGF)-β (860206, R&D systems, Minneapolis, MN).

After staining, cells were washed with permeabilization/wash buffer then resuspended in Cell Staining Buffer. Samples were analyzed by 4-laser Aurora spectral flow cytometer (Cytek, Fremont, CA). The acquired raw data from the spectral flow cytometer were unmixed by SpectroFlo software (Cytek). Unmixed data was curated and analyzed with FlowJo 10.8 software (BD Biosciences).

Double Negative T Cell Sorting

DN T cells were isolated from gld mice lymph nodes. Briefly, single-cell suspension of lymphocytes from lymph nodes was preincubated with anti-CD16/CD32 Fc block (S17011E, BioLegend) and stained in Cell Staining Buffer (BioLegend) with fluorochrome-labeled antibodies: APC-Cy7 anti-CD45 (30-F11), BV421 anti-TCRβ (H57-957), CD4 Alexa Fluor 488 (GK1.5), and CD8 PE (53-6.7) from BioLegend. After staining, cells were washed and resuspended with BD FACS Pre-Sort Buffer (BD Biosciences). Cells were stained with propidium iodide (PI) (Thermo Fisher Scientific) right before sorting for viability assay. PI⁻ TCRβ⁺ CD4⁻ CD8⁻ cells were sorted using the MoFlo Legacy or the XDP cell sorter (Beckman Coulter, Brea, CA).

Treg and CD4⁺ T Cell Isolation

Tregs and CD4⁺ T cells were isolated from WT mouse spleens using CD4⁺ CD25⁺ Regulatory T Cell Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s protocol. The final eluted fraction containing CD4⁺ CD25⁺ cells was used as Treg transfer for both in vitro and in vivo studies. We confirmed that these cells were FoxP3⁺ CD4⁺ Tregs by flow cytometry analysis before performing adoptive transfer (Supplemental Fig. S1). The flow-through fraction containing CD4⁺ CD25^- T cells was used for CD4⁺ T cell transfer (negative control) for the in vitro scratch experiment.

In Vitro Coculture of Renal Tubular Epithelial Cells and T Cells to Quantify Repair after Injury

The Boston University mouse proximal tubular (BUMPT-306) epithelial cells (PTECs) were cultured with DMEM with 5% FBS and 100 U/mL penicillin and streptomycin in a 96-well plate (Sartorius, Niedersachsen, Germany). Cultured cells were exposed to hypoxic condition (1% O₂) according to the following protocol. For hypoxia induction, culture plates were placed in a modular incubator chamber, and the chamber was flushed with gas mixture containing 1% O₂, 5% CO₂, and 94% N₂ for 3 min. Subsequently, the chamber was completely sealed and placed into a cell culture incubator for 12 h.

To mimic PTEC repair in vitro, we used a scratch wound assay, a technique used for studying the effects of cell-cell interactions on cell migration (27). Homogenous scratch wounds were created to the PTEC monolayer using a 96-pin WoundMaker Tool (Sartorius). Immediately after scratch, activated DN T cells, CD4⁺ Tregs, and CD4⁺ T cells were transferred at a 5:1 ratio. Cells were incubated under normoxia and the phase-contrast images were acquired every 2 h using a real-time cell analysis system (Incuyte Live-Cell Analysis, Sartorius). The wound closure was quantified from time-lapse phase images, and values were expressed as the relative wound density (Incucyte Scratch Wound Analysis Software Module, Sartorius).

T Cell Adoptive Transfer

Vehicle, DN T cells, or Tregs were injected twice at 6 h and 48 h after reperfusion via retro-orbital intravenous injection. PBS was used as the vehicle, and isolated cells were resuspended in PBS right before the injection. DN T cells (5 × 10⁶) and 1 × 10⁶ Tregs were injected into each mouse per injection.

Quantification of mRNA by Real-Time Quantitative RT-PCR

After collecting whole kidneys, the upper one-third part of each kidney, which includes the cortex and medulla, was immediately immersed into RNAlater (Thermo Fisher Scientific). These tissues were subsequently used for RNA isolation followed by RT-PCR. The remaining kidney tissues were used for lymphocyte isolation for flow cytometry or histologic evaluation. Total RNA was extracted from kidney tissue with RNeasy Mini kit (Qiagen, Valencia, CA) and reverse-transcribed using high-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA). Real-time PCR was performed in QuanStudio 12 Flex (Applied Biosystems) using the PowerUp SYBR Green Master mix (Applied Biosystems) for detection of mRNA expression of Tgfb1, Acta2 (encoding α-smooth muscle actin, α-SMA), Col1a1, and Col4a1. Gapdh gene expression was used as the internal control. Relative fold expression values were calculated with a ΔΔ cycle threshold method. The primer sequences for each gene are provided in Supplemental Table S1.

Statistics

Data are expressed as means ± SE. Two group means were compared with a two-tailed t test. Three or more group means were compared using one-way ANOVA followed by Tukey’s post hoc analyses. Kruskal–Wallis test followed by Dunn’s test was used for non-normally distributed variables. All statistical analyses were performed using GraphPad Prism version 10 (GraphPad Software, La Jolla, CA). P values of <0.05 were considered statistically significant.

RESULTS

Kidney DN T Cells Decreased After Severe Ischemic AKI

Severe unilateral ischemic AKI was induced in mice then DN T cells and other conventional T cell subsets were evaluated at 1 wk and 3 wk after IRI. The gating strategies for kidney T cells are provided in Supplemental Fig. S2. DN T cells significantly decreased during recovery phase compared with steady-state level by both percentages (Steady state 18.0 ± 0.5% of αβ T cells; 1 wk 10.1 ± 0.7%, P < 0.001; 3 wk 6.1 ± 0.4%, P < 0.001) and numbers (3.9 ± 0.2 × 10⁵; 2.8 ± 0.2 × 10⁵, P = 0.006; 1.2 ± 0.2 × 10⁵, P < 0.001), whereas CD4⁺ Tregs substantially increased (1.5 ± 0.2%; 7.0 ± 0.3%, P < 0.001; 14.6 ± 1.1%, P < 0.001) (1.8 ± 0.2 × 10⁴; 10.1 ± 1.0 × 10⁴, P < 0.001; 15.2 ± 2.5 × 10⁴, P < 0.001) (Fig. 1). Percentages of CD8⁺ T cells increased at 1 wk and 3 wk, whereas CD4⁺ T cells remained unchanged (Supplemental Fig. S3A). CD8⁺ Tregs were also studied given the emerging interest as a potential regulatory cell (28–30). CD8 Tregs (CD8⁺ Ly49⁺ CD122⁺) increased at 1 wk and then decreased at 3 wk (Supplemental Fig. S3B).

Figure 1.

Changes in anti-inflammatory T cells during repair phase after severe acute kidney injury (AKI). Double-negative (DN) T cells significantly decreased after severe ischemic AKI by percentages and numbers, whereas CD4⁺ regulatory T cells (Tregs) markedly increased. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc analysis (n = 8 for steady-state kidneys, n = 10 for 1 wk, n = 11 for 3 wk). Data are from three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001. The steady-state group refers to uninjured normal kidneys from naïve control mice. IRI, ischemia-reperfusion injury.

Activation and Proliferation of DN T Cells During Repair After Severe AKI

To study activation and proliferation of DN T cells during AKI recovery, we measured CD44, CD62L, CD69, and Ki67 expression in DN T cells. Effector memory (EM) phenotype (CD44^hi CD62L^lo) DN T cells decreased (steady state 92.2 ± 0.5%; 1 wk 70.2 ± 1.1%, P < 0.001; 3 wk 81.3 ± 2.8%, P < 0.001), whereas central memory (CM) phenotypes (CD44^hi CD62L^hi) increased (6.5 ± 0.4%; 28.8 ± 1.1%, P < 0.001; 13.1 ± 2.4%, P = 0.038) in DN T cells during recovery phase. Activation marker CD69 was downregulated (99.2 ± 0.1%; 95.9 ± 0.6%, P = 0.016; 85.5 ± 1.0%, P < 0.001). DN T cell proliferation decreased during recovery phase (Ki67, 97.8 ± 0.5%; 95.1 ± 0.9%, P = 0.133; 90.5 ± 1.2%, P < 0.001) (Fig. 2).

Figure 2.

Double-negative (DN) T cell activation and proliferation during repair phase after severe acute kidney injury (AKI). Effector-memory (EM) (CD44^hi CD62L^lo) DN T cells decreased significantly after severe AKI during repair, whereas percentages of central-memory (CM) (CD44^hi CD62L^hi) DN T cells increased. Expression of markers for activation (CD69) and proliferation (Ki67) decreased significantly in DN T cells during repair phase. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc analysis (n = 8 for steady-state kidneys, n = 10 for 1 wk, n = 11 for 3 wk). Data are from three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001. The steady-state group refers to uninjured normal kidneys from naïve control mice. IRI, ischemia-reperfusion injury.

Changes in DN T Cell Immune Checkpoint Molecules PD1 and TIGIT as Well as NK1.1 During Repair After Severe AKI

NK1.1⁺ and PD1⁺ DN T cells are known to be two major subsets of kidney DN T cells (31). NK1.1 expression (33.4 ± 2.9%; 53.3 ± 1.6%, P < 0.001, 50.4 ± 2.0, P < 0.001) and PD1 expression (1.4 ± 0.1%; 1.8 ± 0.2%, P = 0.820; 3.3 ± 0.7, P = 0.020) were upregulated at 3 wk by percentages. However, the absolute numbers of the DN T cells with positive expression of NK1.1 or PD1 did not increase due to decreased total DN T cell numbers. A newly recognized immune checkpoint molecule with a role in early AKI, TIGIT, was also measured (32), and we found TIGIT expression in DN T cells was decreased (1.9 ± 0.2%; 2.3 ± 0.3%, P = 0.461; 0.9 ± 0.2, P = 0.010) at 3 wk after AKI (Fig. 3).

Figure 3.

Changes in NK1.1 and immune checkpoint molecules in double-negative (DN) T cells in the repair phase after severe acute kidney injury (AKI). NK1.1 and PD1 expression was upregulated at 3 wk by percentages, whereas T-cell immunoreceptor with Ig and ITIM domains (TIGIT) expression was decreased at 3 wk after AKI. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc analysis (n = 8 for steady-state kidneys, n = 10 for 1 wk, n = 11 for 3 wk). Data are from three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001. The steady-state group refers to uninjured normal kidneys from naïve control mice. IRI, ischemia-reperfusion injury.

DN T Cell Phenotypes Were Distinct From CD4⁺ and CD8⁺ T Cells After Severe AKI

Immunologic phenotypes of DN T cells were compared with conventional CD4⁺ and CD8⁺ T cells 3 wk after severe AKI. DN T cells had more central-memory (CD44^hi CD62L^hi) phenotype compared with CD4⁺ and CD8⁺ T cells (DN, 13.1 ± 2.4%; CD4, 2.4 ± 0.2% P < 0.001; CD8, 1.8 ± 0.1%, P < 0.001) (Fig. 4A). When CD44 and CD62L expression was compared with the contralateral normal kidneys, DN T cells from post-AKI kidneys showed lower effector-memory (CD44^hi CD62L^lo) and higher central-memory (CD44^hi CD62L^hi) subtypes compared with those from the nonischemic kidneys, whereas CD4⁺ and CD8⁺ T cells exhibited opposite trends (Fig. 4B). In nonischemic kidneys, there were substantial percentages of naïve phenotypes in CD4⁺ and CD8⁺ T cells, but not in DN T cells. There were few naïve phenotypes in postischemic kidneys in DN, CD4⁺, and CD8⁺ T cells (Fig. 4B).

Figure 4.

Phenotypical differences between double-negative (DN) T cells, CD4⁺ and CD8⁺ T cells 3 wk after severe acute kidney injury (AKI). A: DN T cell CD44 and CD62L expression was compared with CD4⁺ and CD8⁺ T cells at 3 wk after severe ischemic AKI. DN T cells had a lower effector-memory (EM) phenotype (CD44^hi CD62L^lo) compared with CD4⁺ T cells during post-AKI repair. There were few central-memory (CM) phenotypes (CD44^hi CD62L^hi) in CD4⁺ and CD8⁺ T cells. B: when CD44 and CD62L expression was compared with the contralateral normal kidneys, DN T cells from post-AKI kidneys had lower effector-memory and higher central-memory phenotypes compared with those from the contralateral normal kidneys. CD4⁺ and CD8⁺ T cells showed opposite trends. There were few naïve phenotype T cells in post-AKI kidneys. Only CD4⁺ and CD8⁺ T cells had substantial naïve phenotypes in normal kidneys, and they decreased in the post-AKI kidneys. C: Ki67 expression among three different T cell subsets was compared in post-AKI kidneys at 3 wk. CD8⁺ T cells had the highest expression of CD69. Ki67 expression was lower in the DN T cells compared with CD4⁺ and CD8⁺ T cells. D: when CD69 and Ki67 expression was compared with the corresponding subsets from the contralateral normal kidneys, CD69 was lower in the DN T cells from post-AKI kidneys than those from contralateral normal kidneys. Ki69 was higher in CD4⁺ and CD8 T⁺ cells from the post-AKI kidneys, whereas DN T cell Ki67 expression remained unchanged. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc analysis (n = 11). Data are from three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001. The open bars and gray bars represent T cells from ischemic and nonischemic (uninjured contralateral) kidneys, respectively.

DN T cell Ki67 expression was lower compared with CD4⁺ and CD8⁺ T cells during post-AKI repair (90.8 ± 1.1%; 94.4 ± 0.3%, P = 0.001; 95.0 ± 0.6%, P = 0.001) (Fig. 4C). CD69 expression in DN T cells from postischemic kidneys was lower compared with those of nonischemic kidneys whereas CD8⁺ T cells showed the opposite trend. Ki67 expression was lower in CD4⁺ and CD8⁺ T cells from nonischemic kidneys compared with those from ischemic kidneys, whereas it was consistently high in both nonischemic and postischemic kidneys in DN T cells (Fig. 4D).

DN T cell NK1.1 expression was substantially higher than CD4⁺ and CD8⁺ T cells, and PD1 and TIGIT expression was lower than CD4⁺ T cells (Supplemental Fig. S4).

Cytokine Production by DN T Cells After Severe AKI

Intracellular cytokine analysis of DN T cells was performed after severe AKI. After 3 wk from severe AKI, DN T cells from postischemic kidneys had significantly lower expression of inflammatory cytokines including IL-17A, IFN-γ, and TNF-α, compared with CD4⁺ and CD8⁺ T cells. Profibrotic cytokine TGF-β1 expression was also lower in DN T cells than in CD4⁺ and CD8⁺ T cells (Fig. 5). When cytokine expression was compared with the contralateral nonischemic kidneys, there were no significant changes in proinflammatory cytokines in the DN T cells, whereas there was upregulation of IL-17A, IFN-γ, and TNF-α in the CD4⁺ or CD8⁺ T cells. TGF-β1 was upregulated in DN T cells from postischemic kidneys than those from nonischemic kidneys. IL-10 remained unchanged after severe AKI (Fig. 5). When IL-10 and TGF-β1 expression was compared between DN T cells and Tregs from postischemic kidneys, IL-10 was lower in DN T cells whereas TGF-β1 was higher (Supplemental Fig. S5).

Figure 5.

Cytokine production by double-negative (DN) T cells during repair phase after severe acute kidney injury (AKI). Expression of inflammatory cytokines, interleukin (IL)-17A, interferon (IFN)-γ, and tumor necrosis factor (TNF)-α, was significantly lower in DN T cells than CD4⁺ and CD8⁺ T cells 3 wk after severe AKI. TGF-β1 was lower in DN T cells compared with CD4⁺ and CD8⁺ T cells. Compared with the corresponding subsets from the contralateral nonischemic kidneys, IL-17A increased in CD4⁺ T cells, and IFN-γ and TNF-α increased in CD8⁺ T cells, whereas IL-17, IFN-γ, and TNF-α remained unchanged in DN T cells. Transforming growth factor (TGF)-β1 expression was higher in all T cell subsets from postischemic kidneys compared with those from nonischemic kidneys. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc analysis (n = 16). Data are from three independent experiments. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. The open bars and gray bars represent T cells from ischemic and nonischemic kidneys, respectively.

DN T Cells Accelerated Proximal Tubular Epithelial Cell Repair In Vitro

To assess the role of DN T cells in post-AKI repair in vitro, we tested the effect of DN T cell coculture with tubular epithelial cells. PTECs were cultured and then exposed to hypoxia followed by normoxia to model ischemia-reperfusion injury in vitro. After creating consistently sized scratch wounds to tubular epithelial cell monolayer (27), DN T cells were transferred. DN T cells isolated from gld donors were used as commonly performed since large numbers of cells are required (19, 21, 33). gld DN T cells are known to be similar to kidney DN T cells, and they are capable to suppress T cell proliferation in vitro (33). Tregs were used as a positive control, and CD25⁻ CD4⁺ T cells were used as an additional negative control to exclude nonspecific effect of coculturing with lymphocytes. Kinetic quantification was performed with live-cell analysis system. Cellular density at the wound significantly increased in tubular cells treated with DN T cells or Tregs, whereas CD4⁺ T cell-treated cells did not, compared with the vehicle-treated group (Fig. 6).

Figure 6.

Double-negative (DN) T cell coculture accelerated kidney tubular epithelial cell repair in vitro. Scratch wounds were created to proximal tubular epithelial cell monolayer after hypoxia exposure. Activated DN T cells, regulatory T cells (Tregs) (positive control), CD4⁺ T cells (negative control), or vehicle (negative control) were transferred after scratch. Phase-contrast images were acquired serially using a live-cell analysis system. A: representative wound images at 24 h with segmentation lines shown for the initial scratch (blue) and wound (yellow). B: kinetic quantification showing percent wound density for DN, Tregs, CD4⁺ T cells, vehicle, and normoxia control groups. The normoxia control group showed higher wound density compared with the other groups exposed to hypoxia (P < 0.05 for all time points). DN T cell coculture exhibited better wound repair compared with the vehicle and CD4⁺ T cell coculture groups (at 24 h, vehicle 56.3 ± 2.1%; CD4 56.1 ± 1.2%, vs. vehicle P > 0.999; DN 67.5 ± 2.3%, vs. vehicle P = 0.007, vs. CD4 P = 0.006; Tregs 62.1 ± 3.0%, vs. vehicle, P = 0.275, vs. CD4, P = 0.250). n = 8 replicates/group. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc analysis. *P < 0.05 compared with the vehicle control group.

Adoptive Transfer of DN T Cells After Severe AKI Increased GFR and Decreased Kidney Fibrosis

Given the significant decline of DN T cell numbers in post-severe AKI kidneys and in vitro ability of DN T cells to enhance tubular epithelial cell repair, we performed in vivo adoptive transfer experiments to evaluate whether the DN T cell replenishment could reduce kidney fibrosis and enhance renal recovery. DN T cells isolated from gld donors were adoptively transferred at 6 h and 48 h after reperfusion. We previously demonstrated that adoptively transferred gld DN T cells migrate into kidneys (19). CD4 Tregs were also transferred as a positive control. Mice underwent serial GFR measurements during 3-wk follow-up. Fibrosis was measured from kidney histology sections, and profibrotic gene expression was quantified at 3 wk after AKI. Kidney T cells were also studied using flow cytometry (Fig. 7A). The gating strategy for DN T cell sorting is provided in Supplemental Fig. S6.

Figure 7.

Double-negative (DN) T cell treatment after severe acute kidney injury (AKI) increased glomerular filtration rate (GFR) and reduced kidney fibrosis. A: schematic of experimental design. DN T cells, Vehicle (negative control), or regulatory T cells (Tregs) (positive control) were adoptively transferred after severe ischemic AKI. GFR was measured with fluorescein isothiocyanate (FITC)-sinistrin-based method. Kidney sections were stained with Masson’s trichrome to assess kidney fibrosis. Fibrosis genes were quantified with quantitative RT-PCR. Kidney T cells were isolated and studied by spectral flow cytometry. B: GFRs at 3 wk were higher in the DN T cell (n = 17) and Treg (n = 14) transfer groups than the vehicle control group (n = 17). C: Masson’s trichrome staining at 3 wk after severe AKI. Kidney fibrosis in outer medullar was significantly lower in the DN T cell-treated group than the vehicle control group (n = 9/group). D: expression of genes encoding transforming growth factor (TGF)-β and α-smooth muscle actin (αSMA) was lower in the DN T cell-treated group (n =14–17/group). Data are from three independent experiments. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc analysis except for GFR data. The Kruskal–Wallis test followed by Dunn’s test was used for GFR due to non-normal distribution. *P < 0.05; **P < 0.01; ***P < 0.001.

We found significant increases in GFR after 3 wk from ischemic AKI in DN T cell transferred groups and Treg transferred group (vehicle 976.7 ± 29.0 μL/min/100 g; DN T cells 1,092.0 ± 31.4 μL/min/100 g, P = 0.012; Tregs 1,104.9 ± 42.5 μL/min/100 g, P = 0.021) (Fig. 7B). Mice treated with DN T cells exhibited reduced medullary fibrosis (vehicle 76.3 ± 3.2%; DN T cells, 58.1 ± 3.3%, P = 0.008; Tregs 48.7 ± 5.7%, P < 0.001) compared with the vehicle control group. There were no significant differences in cortex (vehicle 57.4 ± 4.5%; DN T cells 45.8 ± 2.6%, P = 0.176; Tregs 41.4 ± 6.0%, P = 0.051) (Fig. 7C).

We found that kidney expression of profibrotic genes Tgfβ (vehicle 1.46 ± 0.14; DN T cells 1.02 ± 0.07, P = 0.011; Tregs 0.92 ± 0.10, P = 0.003) and Acta2 (encoding α-SMA) (vehicle 1.66 ± 0.15; DN T cells 1.22 ± 0.09, P = 0.033; Tregs 1.13 ± 0.12, P = 0.014) were reduced in the DN T cell-treated group and Treg-treated group. Col1α1 was decreased in the Treg transfer group (vehicle 1.56 ± 0.15; DN T cells 1.24 ± 0.15, P = 0.190; Tregs 1.07 ± 0.12, P = 0.043). Col4a1 was comparable between groups (vehicle 1.53 ± 0.14; DN T cells 1.36 ± 0.10, P = 0.527; Tregs 1.30 ± 0.11, P = 0.381) (Fig. 7D).

DN T Cells Decrease Kidney Effector T Cells and Their IL-17A Expression

To begin to elucidate mechanisms by which DN T cells improved GFR and decreased kidney fibrosis after severe AKI, we studied kidney T cells from postischemic kidneys at 3 wk. Although the proportions and numbers of CD4⁺, CD8⁺, and DN T cells were comparable between the groups, we found percentages and numbers of effector-memory CD4⁺ T cells were lower in the DN T cell and Treg-treated groups compared with the vehicle-treated group (vehicle 96.9 ± 0.3%; DN T cells 94.7 ± 0.2%, P < 0.001; Tregs 93.7 ± 0.3%, P = 0.021) (Fig. 8A). We also measured cytokine expression on kidney CD4⁺ T cells and found that IL-17A expression was reduced in DN T cell and Treg-treated groups (vehicle 7.0 ± 1.3%; 3.0 ± 0.4%, P = 0.008; 2.8 ± 0.5, P = 0.020) (Fig. 8B). The other cytokines remain unchanged. Thus, a potential mechanism by which DN T cell reduces kidney fibrosis through decreased effector T cells and IL-17A.

Figure 8.

Double-negative (DN) T cell adoptive transfer after severe acute kidney injury (AKI) decreased effector CD4⁺ T cells and interleukin (IL)-17 production. T cells from post-AKI kidneys were analyzed at 3 wk after severe AKI followed by adoptive transfers. A: effector-memory phenotype CD4⁺ T cells were lower in the DN T cell and Treg transfer group. n =10–11/group. Data are from two independent experiments. B: IL-17A expression in CD4⁺ T cells was lower in the DN T cell (n = 15) and regulatory T cells (Treg)-treated mice (n = 9) than in the vehicle control group (n = 15). Data are from three independent experiments. Statistical analyses were performed using one-way ANOVA followed by Tukey’s post hoc analysis. *P < 0.05; **P < 0.01; ***P < 0.001. CM, central-memory; EM, effector-memory.

DISCUSSION

Based on previous data that Tregs mediate organ repair (34) and emerging data on the important role of kidney DN T cells in prevention of AKI (19, 21, 31), we hypothesized that DN T cells undergo functional changes during recovery and could have a reparative role after severe ischemic AKI. Analysis of kidney T cells after severe ischemic AKI using tissue digestion and spectral flow cytometry demonstrated that DN T cells decrease substantially during the repair phase and undergo long-term distinct immunologic changes compared with kidney CD4⁺, CD8⁺, and CD4⁺ Tregs. DN T cell coculture enhanced tubular epithelial cell repair following injury. Addition of DN T cells to injured renal tubular epithelial cells in vitro accelerated wound repair comparable with the “gold” standard Tregs. Adoptive transfer of DN T cells administered after severe AKI increased GFR and reduced kidney fibrosis, as well as reduced effector-memory CD4⁺ T cells and IL-17A expression.

DN T cells are rare in lymphoid organs and peripheral blood, but they constitute significant proportions of kidney TCRαβ⁺ T cells both in mice and human kidneys (17–20). It was thought that DN T cells originated from CD4⁺ or CD8⁺ T cells by loss of receptors, however kidney TCRαβ⁺ DN T cells were identified in mice lacking CD4⁺ T cells (MHC II-deficient mice) and those lacking CD8⁺ T cells (β2m-deficient mice), which suggest DN T cells can be derived from distinct progenitors (31). Kidney DN T cells expand early after moderate AKI kidneys within the first 24 h, and then rapidly decrease at 72 h below the steady-state level (19). In the present study, we found that kidney DN T cells significantly decrease both in numbers and proportion at later time points until 3 wk after severe AKI. Interestingly, DN T cell kinetics during the AKI repair were opposite to CD4⁺ Tregs that showed marked expansion (15, 16, 19, 31).

Though we studied Tregs as a “positive control” anti-inflammatory cell, we found a significant expansion of CD8⁺ Tregs (CD8⁺ CD122⁺ Ly49⁺), a relatively understudied T cell subset, during repair after severe AKI. The expansion of CD8⁺ Tregs has also been demonstrated by a recent study using a cerebral IRI model (35). Emerging evidence suggests that CD8⁺ CD122⁺ Tregs are involved in immune regulation as their counterpart, CD4⁺ Tregs (28, 29). Given the recent promising data showing the protective role against other organ IRI (35), CD8⁺ Tregs’ role in AKI is a promising topic of future study.

Previous studies have shown that DN T cells had distinct immunophenotypical features during early injury compared with kidney CD4⁺ and CD8⁺ T cells (19, 31). In the present study, we observed late immunologic changes toward more central-memory phenotypes with less proliferation rather than effector-memory phenotypes, an opposing trend with CD4⁺ and CD8⁺ T cells during post-AKI recovery. These long-term changes and distinct kinetics/phenotype of DN T cells, compared with Tregs, CD4⁺, and CD8⁺ cells, compose the complex immune response to repair and fibrosis after severe AKI.

Consistent with our earlier findings showing that PD1⁺ subset is a potent early responder in ischemic AKI (31), we observed increased expression of the PD1⁺ during the recovery phase after severe AKI as well. However, the absolute number of PD1⁺ DN T remained comparable with normal kidneys due to the substantial decline in the total number of DN T cells despite PD1 upregulation. Besides PD1 expression, we also studied a novel immune checkpoint molecule, TIGIT, given the recent data on its importance on AKI (32). In contrast to upregulation of TIGIT during the early injury phase (32), TIGIT⁺ DN T cells decreased during the recovery phase. Considering the important role of immune checkpoint molecules in AKI (32, 36, 37) and continued clinical experience of checkpoint inhibitor-associated AKI in patients with cancer (38), we believe that findings from our current study have potential clinical relevance.

Proximal tubular epithelial cells are a primary target of ischemic injury and previous studies have demonstrated direct interactions between these cells and T cells in vitro (11). We therefore investigated the reparative capacity of DN T cells by coculturing them with wounded PTECs. DN T cells significantly enhanced wound healing capacity of PTECs following hypoxic injury. Thus, there could be DN T cell-driven reparative mediators, affecting tubular epithelial cells, which needs to be further explored.

Although the immunologic mechanisms of long-term injury after AKI are less understood than early injury, chronic T cell expansion is frequently seen during AKI repair and fibrosis (12, 13). More specifically, CD4⁺ T cells skewed toward effector-memory phenotype with upregulated activation markers (12, 13). Taken together with the findings from other organ fibrosis models (39, 40), chronic proinflammatory T cell activation is likely to have an important role in kidney fibrosis and CKD transition. Inhibition of late CD4⁺ T cell activation could be a potential therapeutic target to reduce kidney fibrosis and CKD transition. In the present study, we observed that post-AKI DN T cell repletion reduced effector-memory CD4⁺ T cells, consistent with previously found CD4⁺ T cell suppressive function (19). Another important finding was that DN T cells reduced IL-17 expression in CD4⁺ T cells. IL-17 and IL-17-producing T cells were suggested as important contributors to AKI and CKD transition (41–43). There was an increase of IL-17 expressing T cells during post-AKI repair, and genes related to IL-17 pathway were upregulated in kidney fibrosis (14). Elevated IL-17 expression was observed in kidney biopsy specimens from patients with kidney fibrosis (44). Blocking IL-17⁺ cell activation or deletion of Il17 gene mitigated renal injury (41–43). Taken together, the reparative effect of DN T cells could be attributed to inhibiting effector CD4⁺ T cells and IL-17 downregulation. Since the IL-17 downregulation was observed in CD4⁺ Treg-treated group as well, the IL-17-dependent mechanism could also be involved in the well-known protective effect of Treg-based therapy (45).

The immunoregulatory role of DN T cells has also been shown in other chronic disease models such as allograft rejection, graft-versus-host disease, and type 1 diabetes (46). However, DN T cells’ proinflammatory pathogenic roles have also been reported particularly in autoimmune diseases such as lupus, psoriasis, and Sjogren syndrome (47). These discordant findings are likely due to their heterogeneity or plasticity (48). For example, a recent study using single-cell RNA sequencing analysis proposed five different subsets of DN T cells (49). Discovering a specific marker for reparative DN T cell subset is needed for a deeper understanding of this population.

The current study has several limitations. First, although we focused on T cell-mediated mechanisms in AKI repair, other types of kidney immune cells such as macrophages, dendritic cells, neutrophils, B cells, and innate lymphoid cells could also play important roles in AKI repair (50–52). Thus, we cannot rule out a collateral effect of DN T cells on other types of kidney immune cells. Second, there are currently no available DN T cell-specific markers, thus we were unable to deplete kidney DN T cells or use models lacking DN T cells. The discovery of reparative kidney DN T cell-specific markers is warranted for future studies to use these techniques to further understand the role of DN T cells. Another limitation was that we used gld DN T cells, which are not identical to WT DN T cells. However, since DN T cells are rarely present in lymphoid organs, it is challenging to get enough DN T cells for DN T cell adoptive transfer. We previously demonstrated that adoptively transferred gld DN T cells migrate into post-AKI kidneys (19).

Despite these limitations, our study is novel and has important pathophysiologic and therapeutic implications. The decreasing kinetics of DN T cells could be a potential target involved in kidney fibrosis in other diseases. Cell immunotherapy is an increasing reality, particularly for patients with cancer. The regulatory cell-based therapy field has flourished with recent ongoing clinical trials in kidney diseases (53). The present study identifies a novel therapeutic approach to administer expanded DN T cells to accelerate recovery and decrease fibrosis in patients after severe AKI.

DATA AVAILABILITY

The data that support the findings of this study are available from the corresponding author upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Figs. S1–S6 and Supplemental Table S1: https://doi.org/10.6084/m9.figshare.24547087.

GRANTS

K.L. was supported by Korea Health Industry Development Institute Grant HI19C1337, National Research Foundation of Korea Grant NRF-2021R1A6A3A03039863, Samsung Medical Center Grant SMO1230251, and the Young Investigator Research Grant from the Korean Nephrology Research Foundation (2023). J.T.K. was supported by Dr. Werner Jackstädt Foundation Award S 134-10.117. A.M.N.-R. was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) PAR-21-071 Diversity Research Supplement under Grant 3R01DK123342-03S1. S.N. was supported by a Carl W. Gottschalk Research Scholar Grant from the American Society of Nephrology under Grant 134535, the Edward S. Kraus award from the Johns Hopkins School of Medicine Division of Nephrology, National Kidney Foundation Serving Maryland and Delaware mini-grant 142076, and NIDDK Grants R01DK132278, R01DK123342, and R01DK104662. H.R. was supported by NIDDK Grants R01DK104662 and R01DK123342.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

H.R. conceived and designed research; K.L., S.G., J.T.K., A.M.N.-R., and S.N. performed experiments; K.L., S.N., and H.R. analyzed data; K.L., L.J.A., S.N., and H.R. interpreted results of experiments; K.L. and L.J.A. prepared figures; K.L. and H.R. drafted manuscript; K.L., S.G., J.T.K., A.M.N.-R., L.J.A., S.N., and H.R. edited and revised manuscript; K.L., S.G., J.T.K., A.M.N.-R., L.J.A., S.N., and H.R. approved final version of manuscript.