The mitochondrial fission protein DRP1 influences memory CD8+ T cell formation and function

Abstract

Pharmacological methods for promoting mitochondrial elongation suggest that effector T cells can be altered to support a memory T cell–like metabolic state. Such mitochondrial elongation approaches may enhance the development of immunological memory. Therefore, we hypothesized that deletion of the mitochondrial fission protein dynamin-related protein 1 (DRP1) would lead to mitochondrial elongation and generate a large memory T cell population, an approach that could be exploited to enhance vaccination protocols. We find that, as expected, while deletion of DRP1 from T cells in dLckCre × Drp1^flfl does compromise the magnitude and functionality of primary effector CD8+ T cells, a disproportionately large pool of memory CD8+ T cells does form. In contrast to primary effector CD8+ T cells, DRP1-deficient memory dLckCre × Drp1^flfl CD8+ T cells mount a secondary response comparable to control memory T cells with respect to kinetics, magnitude, and effector capabilities. Interestingly, the relative propensity to form memory cells in the absence of DRP1 was associated with neither differentiation toward more memory precursor CD8+ T cells nor decreased cellular death of effector T cells. Instead, the tendency to form memory CD8+ T cells in the absence of DRP1 is associated with decreased T cell receptor expression. Remarkably, in a competitive environment with DRP1-replete CD8+ T cells, the absence of DRP1 from CD8+ T cells compromised the generation of primary, memory, and secondary responses, indicating that approaches targeting DRP1 need to be carefully tailored.

The absence of DRP1 limits primary CD8+ T cell responses, yet memory and secondary responses are intact, unless in competition with DRP1-replete CD8+ T cells.

1. Introduction

Immune resolution of an initial encounter with a pathogen or vaccination usually results in populations of memory T cells that are specific for the introduced antigens and can prevent or reduce the severity of a subsequent infection. The majority of effector T cells that initially expand from infection or immunization will undergo cell death during a contraction phase, while a small fraction will differentiate into a lasting memory T cell population.^1–3 These memory T cells exist in larger numbers and respond more rapidly than naïve T cells.⁴ Therefore, the efficacy of vaccines could be improved by supporting the development of memory T cells.

T cells utilize mitochondrial dynamics to respond to their metabolic needs. The size and shape of mitochondria are altered through a balance of fission and fusion.⁵ In response to T cell receptor (TCR) engagement, quiescent T cells can rapidly increase metabolism to support expansion and differentiation. Effector T cells engage both aerobic glycolysis and oxidative phosphorylation (OXPHOS) metabolism to support the energetic and biosynthetic demands of an activated, proliferative cell.^6–8 Furthermore, memory T cells favor fatty acid oxidation and OXPHOS, which are supported by increased mitochondrial mass and an elongated mitochondrial network in memory T cells.^9,10 The elongated mitochondrial morphology and increased mitochondrial mass of memory T cells create a metabolic spare respiratory capacity (SRC) that allows memory cells to rapidly respond to a secondary antigen encounter.¹¹ As the mitochondrial fission process primarily relies on the scission activity of the GTPase, dynamin-related protein 1 (DRP1),^12,13 targeting DRP1 may provide an avenue to augment memory T cell development, survival, or responsiveness.

Supporting the idea that the manipulation of mitochondrial morphology may provide an opportunity to augment immunological memory, the pharmacological treatment of effector T cells with the mitochondrial fusion enforcer M1 and mitochondrial division inhibitor M-divi endows effector T cells with increased SRC and mitochondrial mass comparable to that of in vitro memory-like T cells.¹⁰ At the time of this discovery, M-divi was widely used as a specific inhibitor of the mitochondrial fission protein DRP1, but this has since come into question when it was shown to not affect mitochondrial morphology and to inhibit complex I of the electron transport chain in mammalian cells.¹⁴ Thus, it cannot be concluded from this study that inhibition of mitochondrial fission is directly responsible for promoting memory cell differentiation.

The notion that targeting DRP1 can be leveraged to support memory T cell development and function is potentially countered by several roles for DRP1 in T cell biology. First, deletion of DRP1 in T cells is likely to be detrimental to activation and proliferation. During the process of T cell activation, DRP1 assists in positioning mitochondria near the immune synapse, without which TCR signal strength is reduced.¹⁵ Diminished TCR signaling could limit not only T cell expansion, but also differentiation into effector populations. Additionally, interfering with mitochondrial dynamics can hinder cell division.^16,17 Pertaining to this, Simula et al.¹⁸ previously demonstrated that LckCre × Drp1^flfl mice have reduced numbers of thymocytes and mature T cells in the blood and spleen, and only produce about half as many antigen-specific T cells compared with control mice 4 d after exposure to tumor antigen. Thus, constraining DRP1 may limit the pool of memory precursor cells. Second, DRP1 is also implicated in maintaining the health of mitochondria, as damaged mitochondria are removed through the process of mitophagy.¹⁹ T cells with unhealthy mitochondria might not be expected to efficiently populate the memory pool or might be compromised in their capacity to respond to a secondary exposure to antigen. However, T cells become less susceptible to programmed cell death when DRP1 is inhibited.²⁰ DRP1-dependent fission also occurs following TCR signaling and ultimately results in activation induced cell death.²¹ Thus, DRP1 has complex roles in contributing to T cell survival and memory T cell development that need further clarification before being effectively leveraged for the purposes of improved vaccination.

Based on these considerations, we sought to generate further insights into the relationship between mitochondrial fission and T cell differentiation, particularly how DRP1 affects the development of in vivo effector T cells, memory T cells, or the capacity for those memory cells to mount a secondary response to antigen challenge. We therefore utilized mouse models of T cell–specific DRP1 deletion to address these unknowns.

2. Methods

2.1. Animals

All mice were treated in accordance with policies established by the University of Virginia Animal Care and Use Committee protocol 3292. Drp1^flfl mice were developed by Dr. Hiromi Sesaki and generously provided by Dr. David Kashatus at the University of Virginia.²² E8iCreERT2 mice were generously provided by Dr. Dario Vignali at the University of Pittsburgh.²³ C57BL/6NCr (027) and B6-Ly5.2/Cr (564) mice were purchased from the National Cancer Institute. dLckCre (012837) and Thy1.1 (000406) mice were purchased from the Jackson Laboratory.^24,25

2.2. In vitro T cell stimulation

Mouse T cells were labeled with Cell Trace Violet (CTV) (5 µM; Invitrogen) according to the manufacturer’s protocol. CTV-labeled mouse T cells were then cultured in vitro for 2 to 3 d with anti-CD3 (5 µg/mL plate bound; eBioscience; clone 17A2), anti-CD28 (2 µg/mL soluble; eBioscience; clone 37.51), and interleukin-2 (10 IU/mL; Chiron). Cell culture media is RPMI with 10% fetal bovine serum (Corning), HEPES, nonessential amino acids, essential amino acids, gentamicin, sodium pyruvate, L-glutamine, and β-mercaptoethanol (all from Gibco).

2.3. T cell stimulation for cytokine assessment

Murine T cells were stimulated in vitro with ovalbumin (OVA_257–264) peptide pulsed (10 µg/mL; Genscript) LB15.13 antigen-presenting cells (American Type Culture Collection) or with PMA (0.2 µg/mL; Thermo Fisher) and ionomycin (0.14 µg/mL; Sigma-Aldrich) in the presence of brefeldin A (3 µg/mL; Thermo Fisher) for 4 to 6 h. Cell culture media is RPMI with 10% fetal bovine serum (Corning), HEPES, nonessential amino acids, essential amino acids, gentamicin, sodium pyruvate, L-glutamine, and β-nercaptoethanol (all from Gibco).

2.4. Flow cytometry staining

Single-cell suspensions were prepared via mechanical homogenization, 70 µm filtration (Genesee), and red blood cell lysis (Sigma-Aldrich). Cell viability was assessed with fixable aqua or far-red LIVE/DEAD Fixable Dead Cell Stains (Life Technologies) in phosphate-buffered saline (PBS). H-2K^b/SIINFEKL dextramer (Immudex) was used to identify OVA_257–264–specific T cells followed by a panel of all other surface antibodies. Surface staining was performed in fluorescence-activated cell sorting (FACS) buffer (PBS containing 2% bovine serum albumin, 0.08% sodium azide). For cytokine staining, Cytofix/Cytoperm Fixation and Permeabilization Solution (BD Biosciences) and Perm/Wash Buffer (BD Biosciences) were used. For nuclear staining, Foxp3/Transcription Factor Staining Buffer Set (eBioscience) was used. At the end of staining, samples were preserved using FACS lysis (BD Biosciences) and stored in FACS buffer. Live cell staining was performed entirely in PBS and immediately acquired, and 123count eBeads (Invitrogen) were used to calculate cell counts and UltraComp eBeads (Invitrogen) for single-stain preparations.

2.5. MitoTracker green and TMRE staining

For MitoTracker green (MTG) and TMRE staining, samples were stained with Aqua Live/Dead dye and antibodies specific to surface proteins as above and then incubated at 37 °C for 20 min with 50 nM MTG (M-7514; Thermo Fisher) or 100 nM TMRE (ab113852; Abcam) diluted in serum-free RPMI with HEPES, nonessential amino acids, essential amino acids, gentamicin, sodium pyruvate, L-glutamine, and β-mercaptoethanol (all from Gibco). All samples were washed with 1× PBS twice and immediately analyzed by flow cytometry while keeping them on ice in 0.2% fetal bovine serum in 1× PBS.

2.6. Flow cytometry acquisition and analysis

Samples stained for flow cytometry were collected and compensation algorithm applied on an AttuneNxT cytometer (Thermo Fisher). The resulting FCS files were analyzed, and plots were generated using FlowJo 10 (BD). All flow cytometry data were pregated on single cells, lymphocyte scatter, viable cells, and CD45 expression.

2.7. In vivo priming

Mice were injected intraperitoneally with 100 µg αCD40 (FGK45; Bio X Cell), 75 µg polyI:C (Invivogen), and 500 µg ovalbumin (Sigma-Aldrich) in 200 µL of PBS, ph 7.4 (Gibco).

Secondary responses were initiated in OVA-primed mice by intraperitoneal challenge with 2 × 10⁸ plaque-forming units recombinant adenovirus-expressing OVA, originally provided by Dr. Young Hahn (University of Virginia).

2.8. Seahorse mitochondrial stress test

CD8+ T cells were magnetically enriched (Thermo Fisher) washed with 1× PBS and seeded in XFe96 cell culture miniplate in XF minimal media (Agilent) supplemented with 10 mM glucose (Sigma-Aldrich), 2 mM glutamine (Invitrogen), and 2 mM sodium pyruvate (Thermo Fisher). The oxygen consumption rate (OCR) was measured by extracellular flux analysis using a Seahorse XFe96 analyzer (Agilent) in response to 1 μM oligomycin (Sigma-Aldrich), 1 μM FCCP (Sigma-Aldrich), and 0.5 μM rotenone/antimycin (MP Biomedicals).

2.9. Mixed bone marrow chimera mouse generation

Donor mouse femurs and tibia were collected from Thy1.1+ CD45.2+ wild-type (WT) and Thy1.2+ CD45.2+ dLckCre × Drp1^flfl mice then centrifuged at ≥10,000 g to collect bone marrow. Thy1.2+ CD45.1+ recipient mice were irradiated over 2 sessions for a total of 1300 rads, then 500,000 donor cells were transferred into recipient mice intravenously. Mice were kept on sulfa water for 8 wk prior to experimental use.

2.10. Adoptive T cell transfer

T cells were isolated using magnetic enrichment for CD8+ T cells (Invitrogen) followed by magnetic enrichment for CD44 PE-stained cells (Miltenyi) in magnetic-activated cell sorting buffer (PBS containing bovine serum albumin and EDTA). A total of 10,000 resulting cells were intravenously injected into recipient mice in 100 µL of PBS. Mice were challenged as described above 24 h post–cell transfer.

2.11. Mitochondrial confocal imaging

Naïve or in vitro stimulated CD8+ T cells were incubated with 300 nM MitoTracker Deep Red (Thermo Fisher Scientific) for 30 min, mounted with Wescor Cytopro cytocentrifuge, and then fixed with 4% PFA (Electron Microscopy Sciences) in cytoskeleton buffer. Cells were permeabilized with 0.1% Triton X-100 (Sigma-Aldrich), stained for CD8 (BioLegend; catalog 344702, clone SK1) and DAPI (300 nM; Thermo Fisher Scientific), and mounted with ProLong Gold Antifade reagent (Thermo Fisher Scientific). Secondary donkey anti–mouse IgG Alexa Fluor 488 antibodies (Life Technologies, polyclonal) were used. Images were acquired on a Zeiss LSM880 confocal microscope with Airyscan for super-resolution imaging using a 63×/1.40NA Plan-Apochromat oil objective and 405-, 488-, 561-, and 633-nm lasers and default settings for pinhole and filters. Images were processed using ZEN 2.1 SP1 software (https://hcbi.fas.harvard.edu/resources_software) and FIJI 2.3.0/1.53q (ImageJ; National Institutes of Health). To process mitochondrial images in FIJI, 50-pixel rolling ball background subtraction was used. Temporal color code function in FIJI was used for z-series color coding. Images of MitoTracker Deep Red staining of CD8+ T cells were manually scored in a blinded manner for fissed or fused mitochondria to quantify the mitochondrial morphology of CD8+ T cells.

2.12. Statistical analysis

All data are presented as individual values with mean ± SEM or mean ± SD as indicated in the figure legends. All statistics were calculated by performing unpaired or paired 2-tailed t tests or 1-way analysis of variance for 95% confidence limits in GraphPad Prism 9 (GraphPad Software).

3. Results

3.1. dLckCre × Drp1^flfl mice produce functional effector T cells but not to the extent of Drp1^flfl mice

While previous studies have indicated that the primary effector CD8+ T cell response may be reduced in LckCre × Drp1^flfl mice (where DRP1 is removed prior to maturation of single positive T cells), these studies were performed either in vitro or by interrogating very early (day 4) responses to tumor lysate immunization.¹⁸ Thus, it is unknown whether the magnitude of the primary CD8+ T cells response is compromised in the absence of DRP1 in vivo, nor whether restricting DRP1 expression impacts effector functions or the generation of a secondary response to antigen. Further, the diminution of primary CD8+ T cell responses observed in the prior study could have been a consequence of the loss of DRP1 during T cell development. Therefore, we generated a mouse model of DRP1 restriction in late-development T cells by crossing mice with Cre transgene expression controlled by the distal Lck promoter and Drp1 floxed mice, in which DRP1 is only deleted in the mature CD4+ and CD8+ T cells.^22,24 We characterized the naïve populations of T cells from the thymi and spleens of naïve dLckCre × Drp1^flfl mice (Supplementary Fig. 1). While the subpopulations of thymic T cells are slightly shifted toward single positive CD4+ and CD8+ T cells (Supplementary Fig. 1A–J), ultimately neither the percentage nor the number of mature CD4+ T cells, CD8+ T cells, B cells, and natural killer cells were altered in the spleens of naive dLckCre × Drp1^flfl animals compared with control mice (Supplementary Fig. 1K–V). Furthermore, the mitochondrial mass and mitochondrial membrane potential of naïve CD8+ T cells were not altered in the absence of DRP1 (Supplementary Fig. 1W–Z). Therefore, naïve dLckCre × Drp1^flfl mice have apparently comparable T cell populations to control mice, providing an equivalent baseline population to further investigate the impact of DRP1 loss on T cell activation and differentiation. Given the equivalent mitochondrial membrane potential, we confirmed that loss of DRP1 impacted the fission of mitochondria by stimulating CD8+ T cells in vitro to encourage mitochondrial fission.¹⁰ We visualized mitochondrial morphology by super-resolution confocal microscopy and generated stacked images of individual CD8+ T cells to observe mitochondrial morphology across multiple planes. Following 2 d of stimulation with αCD3 and αCD28, more than 75% of DRP1-deficient CD8+ T cells are dominated by elongated mitochondria compared with 25% of DRP1-replete CD8+ T cells, confirming the impact of the loss of DRP1 on mitochondrial fission (Supplementary Fig. 2). We further assessed these stimulated cells to determine if mitochondrial elongation, due to DRP1 deficiency, enhances the SRC of CD8+ T cells. Surprisingly, DRP1-deficient activated CD8+ T cells did not have significantly altered SRC and perform comparable OXPHOS to DRP1-replete CD8+ T cells as measured by Seahorse analyzer (Supplementary Fig. 3).

We next assayed whether CD8+ T cells become activated, proliferate, and perform effector functions in the absence of DRP1 in vivo. After immunizing dLckCre × Drp1^flfl mice or control mice with agonistic αCD40, polyI:C, and OVA protein, we found that dLckCre × Drp1^flfl mice can generate an OVA_257–264-specific CD8+ T cell response 7 d later (the peak of the primary response) (Fig. 1A–C) but that the magnitude of this response (Fig. 1D) is more than 2.5-fold lower than that found in Drp1^flfl control mice immunized in the same manner. Effector functions were initially assessed after stimulating splenocytes with the mitogens PMA and ionomycin. We found a significant but modest reduction in the proportion of OVA_257–264-specific DRP1-deficient CD8+ T cells that express CD107a, a marker of degranulation (Fig. 1E, F), but with equivalent per-cell expression (Fig. 1G). No differences were found in interferon γ (IFNγ) expression (Fig. 1H–J), while a modest reduction was seen in the proportion of DRP1-deficient OVA_257–264-specific CD8+ T cells to produce the cytokine tumor necrosis factor α (Fig. 1K–L), but again with equivalent per-cell expression (Fig. 1M).

Fig. 1.

Primary effector CD8+ T cell response is diminished in dLckCre × Drp1^flfl mice. Spleens from dLckCre × Drp1^flfl and littermate control mice were collected 7 d postpriming with αCD40, polyI:C, and OVA protein and analyzed by flow cytometry. Representative dot plots of spleen samples from Drp1^flfl (A) and dLckCre × Drp1^flfl (B) mice with percentage (C) and cell number (D) of CD44+ OVA_257–264-specific CD8+ T cells. Spleen samples were restimulated in the presence of brefeldin A with PMA and ionomycin to measure cytokine production in OVA_257–264 multimer + CD8+ T cells. Histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (E) with percentage (F) and fluorescence intensity (G) of CD107a+ of CD44+ OVA_257–264 multimer + CD8+ T cells. Histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (H) with percentage (I) and fluorescence intensity (J) of IFNγ+ of CD44+ OVA_257–264 multimer + CD8+ T cells. Histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (K) with percentage (L) and fluorescence intensity (M) of tumor necrosis factor α+ (TNFα+) of CD44+ OVA_257–264 multimer + CD8+ T cells. (N–S) Spleen samples were restimulated in the presence of brefeldin A with OVA_257–264 peptide pulsed antigen-presenting cells to measure cytokine production. Histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (N) with percentage (O) and fluorescence intensity (P) of CD107a+ of CD44+ OVA_257–264-specific CD8+ T cells. Histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (Q) with percentage (R) and fluorescence intensity (S) of IFNγ+ of CD44+ OVA_257–264 multimer + CD8+ T cells. Individual values and mean ± SEM of n = 10 mice per group shown from 1 of 3 similar experiments. Statistics calculated using unpaired t tests. *P ≤ 0.05; **P ≤ 0.01. GMFI = Geometric mean fluorescence intensity.

However, since DRP1 deletion could compromise TCR signaling and activation,¹⁵ we further questioned if these primary CD8+ T cells were capable of full effector function when stimulated via the TCR.¹⁵ Therefore, we re-stimulated primary effector CD8+ T cells with antigen presenting cells pulsed with SIINFEKL OVA_257–264 peptide. Equivalent percentages of DRP1-deficient and DRP1-replete CD8+ T cells were able to degranulate, as marked by CD107a (Fig. 1N, O); however, there was a notable absence of CD107a high expressors in the DRP1-deficient population (Fig. 1P). As with mitogen stimulation, neither the percentage of cells producing IFNγ (Fig. 1Q, R) nor the amount of IFNγ produced were compromised in the absence of DRP1 (Fig. 1S). We additionally assessed the impacts of DRP1 loss on mitochondrial content. While loss of DRP1 increased elongation of mitochondria, loss of DRP1 alters neither the mitochondrial mass as measured by MTG nor the mitochondrial membrane potential as measured by TMRE (Supplementary Fig. 4A–D), which is consistent with the comparable SRC observed in stimulated CD8+ T cells (Supplementary Fig. 3). Thus, despite the expected impact of loss of DRP1 on CD8+ T cell expansion, a significant primary response develops, and those cells differentiate into effector cells capable of cytokine production.

3.2. DRP1's influence on T cell effector responses is CD8+ T cell intrinsic

Because dLckCre also affects the expression of DRP1 in CD4+ T cells, which are known to support CD8+ T cell responses,²⁶ we determined if the effects of loss of DRP1 on the primary response were due to intrinsic activity of DRP1 within CD8+ T cells. Therefore, we generated E8iERT2Cre × Drp1^flfl mice, which have an inducible CD8+ T cell–specific deletion of DRP1 in response to tamoxifen treatment or sham-treated control mice. Mice were injected intraperitoneally with 75 mg/kg of tamoxifen in corn oil for 3 consecutive days prior to priming. The deletion of DRP1 after tamoxifen treatment was confirmed by Western blot (Supplementary Fig. 5A). We observed a 3-fold decrease in the magnitude of the primary OVA_257–264-specific CD8+ T cell response in tamoxifen-treated mice compared with control mice, indicating that the effects of losing DRP1 on the magnitude of the CD8+ T cell response are intrinsic to CD8+ T cells (Fig. 2A–D). We further determined if these CD8+ T cells, which were activated in the presence of DRP1-replete CD4+ T cells, produce cytokines. We found complete functional responses after stimulating splenocytes with the mitogens PMA and ionomycin (Fig. 2E–2J). However, this method of stimulation for cytokine production does not engage the TCR; therefore, we alternatively restimulated with antigen-presenting cells pulsed with SIINFEKL OVA_257–264 peptide to assess cytokine production. A total of 1.5-fold fewer E8iERT2Cre × Drp1^flfl OVA_257–264-specific CD8+ T cells have CD107a expression (Fig. 2K, L) and about 3-fold fewer produce IFNγ (Fig. 2N, O), while the expression per CD107a+ cell (Fig. 2M) or the amount of IFNγ per cell (Fig. 2P) was not compromised. These results confirm the effect of DRP1 loss on primary CD8+ T cell expansion found in dLckCre × Drp1^flfl CD8+ T cells, indicating that the effects of DRP1 are CD8+ intrinsic and that T cells lacking DRP1 may have a defect in TCR signaling that leads to diminished cytokine production.

Fig. 2.

Diminished primary effector CD8+ T cell response in E8iERT2Cre × Drp1^flfl mice. Spleens were collected from mice 7 d after priming with αCD40, polyI:C, and OVA protein. Representative dot plots of spleen samples from vehicle control (A) and tamoxifen-treated (B) E8iERT2Cre × Drp1^flfl mice with percentage (C) and cell number (D) of CD44+ OVA_257–264 multimer + CD8+ T cells. (E–J) Spleen samples were restimulated in the presence of brefeldin A with PMA and ionomycin to measure cytokine production. Histograms from vehicle control and tamoxifen-treated E8iERT2Cre × Drp1^flfl mice (E) with percentage (F) and fluorescence intensity (G) of CD107a+ of CD44+ OVA_257–264 multimer + CD8+ T cells. Histograms from vehicle control and tamoxifen treated E8iERT2Cre × Drp1^flfl mice (H) with percentage (I) and fluorescence intensity (J) of IFNγ+ of CD44+ OVA_257–264 multimer + CD8+ T cells. (K–P) Spleen samples were restimulated in the presence of brefeldin A with OVA_257–264 peptide pulsed antigen-presenting cells to measure cytokine production. Histograms from vehicle control and tamoxifen-treated E8iERT2Cre × Drp1^flfl mice (K) with percentage (L) and fluorescence intensity (M) of CD107a+ of CD44+ OVA_257–264-specific CD8+ T cells. Histograms from vehicle control and tamoxifen-treated E8iERT2Cre × Drp1^flfl mice (N) with percentage (O) and fluorescence intensity (P) of IFNγ+ of CD44+ OVA_257–264 multimer + CD8+ T cells. Individual values and mean ± SEM of n = 5 mice per group shown from 1 of 2 similar experiments. Statistics calculated using unpaired t tests. *P ≤ 0.05; **P ≤ 0.01. GMFI = Geometric mean fluorescence intensity.

3.3. In vitro stimulated dLckCre × Drp1^flfl cells demonstrate delayed proliferation as well as decreased TCR signaling and expression

To help understand the reduced primary effector CD8+ T cell response found in vivo (Fig. 1D), we questioned whether DRP1-deficient CD8+ T cells failed to become activated, were delayed in proliferation, or were dying. CTV-labeled DRP1-deficient CD8+ T cells, stimulated in vitro with αCD3 and αCD28, demonstrated delayed proliferation compared with control mice. After 2 d of stimulation, DRP1-deficient CD8+ T cells had 1 less division than control CD8+ T cells (Fig. 3A, B). This is consistent with findings from Simula et al.,¹⁸ in which isolated DRP1-deficient T cells stimulated for 5 d in vitro had delayed late-stage proliferation within hours of synchronous release. While dLckCre × Drp1^flfl CD8+ T cells are capable of a surprising amount of division, a greater portion remain in early stages of division, while DRP1-replete T cells progress to higher rounds of division (Fig. 3C). Surprisingly, T cell viability was not compromised during any stage of division within 2 d of stimulation (Fig. 3D). Comparable amounts of CD8+ cells upregulated CD25 (Fig. 3E) and CD44 (Fig. 3F), demonstrating that the expression of activation markers post–TCR engagement was not compromised by deletion of DRP1. A possible explanation for the production of functional effector T cells in the absence of mitochondrial fission would be the upregulation of mitochondrial biogenesis. However, based on PGC1α staining, biogenesis is equivalent in DRP1-deficient and DRP1-replete CD8+ T cells (Fig. 3G). Four days after stimulation, the majority of CD8+ T cells had extensively divided based on CTV dilution, but a small portion of dLckCre × Drp1^flfl CD8+ T cells remained undivided (Fig. 3H, I). Thus, we conclude that DRP1 is not required for CD8+ T cells to become activated and divide, but the kinetics of division are delayed without DRP1.

Fig. 3.

In vitro stimulated dLckCre × Drp1^flfl cells demonstrate delayed proliferation. CTV-stained splenocytes from Drp1^flfl or dLckCre × Drp1^flfl mice were stimulated in vitro with αCD3 and αCD28 in the presence of interleukin-2. Histogram of CTV division 2 d after stimulation (A) and percentage of CD8+ T cells that had divided (B) with proportion of the population (C) and viability of CD8+ T cells in each CTV division (D). Fluorescence intensity of activation markers CD25 (E) and CD44 (F) and mitochondrial biogenesis marker PGC1α (G) on CD8+ T cells per division of CTV. CTV intensity 4 d after stimulation (H) and percent undivided (I) of CD8+ T cells. (J) Percentage of CD8+ T cells from Drp1^flfl or dLckCre × Drp1^flfl mice with phosphorylated CD3ζ at indicated time points after in vitro stimulation with αCD3 normalized to time 0. (K) The expression level (GMFI) of TCRβ on CD8+ T cells from Drp1^flfl or dLckCre × Drp1^flfl naïve mice. Individual values and mean ± SEM of n = 4 mice per group shown from 1 of 2 similar experiments. Statistics calculated using unpaired t tests. *P ≤ 0.05; **P ≤ 0.01; ***P≤0.001; ****P≤0.0001. GMFI = Geometric mean fluorescence intensity.

Taken together with the reduction in the production of some cytokines ex vivo after TCR stimulation, but not after PMA/ionomycin, which bypasses the TCR, we questioned if TCR signaling in DRP1-deficient cells is compromised. Therefore, we stimulated CD8+ T cells with αCD3 over a time course from 1 to 20 min and assessed phospho-tyrosine kinase signaling via flow cytometry. All time points assessed demonstrate a drastic decrease in the phosphorylation of CD3ζ of DRP1-deficient CD8+ T cells starting at a 1.8-fold decrease and becoming more separated over time (Fig. 3J). DRP1 has previously been shown to participate in the structural formation of the immune synapse of Jurkat T cells¹⁵; therefore, aberrations in TCR signaling are likely in the dLckCre × Drp1^flfl model or any model of DRP1 deficiency. We also questioned if DRP1-deficient CD8+ T cells expressed the TCR on the cell surface at levels comparable to DRP1-replete CD8+ T cells. Flow cytometry analysis showed that the expression of the TCRαβ is modestly reduced in DRP1-deficient CD8+ T cells from the spleens of naïve mice compared with DRP1-replete CD8+ T cells (Fig. 3K), which could contribute to the observed reduction in TCR-based signaling. Thus, while the absence of DRP1 does not compromise the ability of the majority of CD8+ T cells to undergo the process of activation and division, TCR signaling is decreased and TCR expression is decreased, potentially reducing the kinetics of T cell activation and division.

3.4. Diminished primary response in dLckCre × Drp1^flfl mice does not interfere with secondary response

After confirming that DRP1-deficient CD8+ T cells can develop into a pool of primary effectors, we next determined how the absence of DRP1 affects memory CD8+ T cell development and their subsequent ability to mount a response to a secondary exposure to OVA antigen. Due to the compromised primary response, on the one hand we anticipated that a secondary response could have similarly reduced magnitude and function. On the other hand, sustained mitochondrial elongation may enhance the generation or activity of memory CD8+ T cells. dLckCre × Drp1^flfl mice were primed with agonistic αCD40, polyI:C, and OVA protein and then challenged 28 d later with recombinant adenovirus-expressing OVA. We analyzed the secondary response 5 d post challenge. Interestingly, we find found an equivalent percentage (Fig. 4A–C) and number (Fig. 4D) of OVA_257–264-specific CD8+ T cells in dLckCre × Drp1^flfl mice and control mice. To quantify cytokine production of OVA_257–264-specific T cells by flow cytometry, we stimulated splenocytes with OVA_257–264 (SIINFEKL) peptide–pulsed antigen-presenting cells in the presence of brefeldin A. In contrast to the primary response, these DRP1-deficient OVA_257–264-specific secondary CD8+ T cells degranulate (Fig. 4E–G), produce IFNγ (Fig. 4H–J), and produce tumor necrosis factor α (Fig. 4K–M) to the same extent as DRP1-replete OVA_257–264-specific secondary CD8+ T cells. Functionally, DRP1 deficient OVA_257–264-specific secondary CD8+ T cells do not exhibit the same defect identified in DRP1-deficient OVA_257–264-specific primary CD8+ T cells.

Fig. 4.

Equivalent secondary response in dLckCre × Drp1^flfl and Drp1^flfl mice. Spleens were collected 5 d after Adeno-OVA challenge of mice that had been initially primed with αCD40, polyI:C, and OVA protein and then rested. Dot plots of spleen samples from Drp1^flfl (A) and dLckCre × Drp1^flfl (B) mice with percentage (C) and number (D) of CD44+ OVA_257–264 multimer + CD8+ T cells. Splenocytes were restimulated in vitro, in the presence of brefeldin A, with OVA_257–264 peptide pulsed antigen-presenting cells to measure cytokine production. Representative histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (E) with percentage (F) and fluorescence intensity (G) of CD107a + of CD44+ OVA_257–264 multimer + CD8+ T cells. Histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (H) with percentage (I) and fluorescence intensity (J) of IFNγ+ of OVA_257–264 multimer + CD8 T cells. Histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (K) with percentage (L) and fluorescence intensity (M) of tumor necrosis factor α+ (TNFα+) of CD44+ OVA_257–264 multimer + CD8+ T cells. Individual values and mean ± SEM of n = 7 mice per group shown from 1 of 3 similar experiments. Statistics calculated using unpaired t tests. GMFI = Geometric mean fluorescence intensity.

3.5. CD8+ memory T cells develop independently of DRP1

Effector CD8+ T cells of a secondary response result from the expansion of a small quiescent memory T cell pool.²⁷ In order to achieve a comparable secondary response to DRP1-replete CD8+ T cells, we hypothesized that DRP1-deficient CD8+ T cells produced memory T cells in greater number or proliferative capacity relative to the magnitude of the primary response. Thus, we enumerated the quiescent memory T cells that developed in the spleen after resting dLckCre × Drp1^flfl and control mice for 28 d postpriming with agonistic αCD40, polyI:C, and OVA protein. The absence of Drp1 in memory CD8+ T cells was confirmed using quantitative polymerase chain reaction (Supplementary Fig. 5B). We found an equivalent number of memory OVA_257–264-specific CD8+ T cells in dLckCre × Drp1^flfl as compared with control mice (Fig. 5A–D). This equivalency in the memory CD8+ T cell population develops despite originating from ∼2.7 times fewer effector CD8+ T cells at 7 d postpriming (Fig. 1D). Memory cells are often found in sites such as lymph nodes and lungs, where they are better poised for rapid secondary responses.²⁸ As DRP1 has previously been shown to be contribute to primary effector CD8+ T cell trafficking,¹⁸ we also evaluated whether the absence of DRP1 affected the distribution and subtypes of memory CD8+ T cells. Interestingly, we identified equivalent amounts of memory OVA_257–264-specific CD8+ T cells in inguinal lymph nodes (Fig. 5E–H) and lungs (Fig. 5M–P). However, some subtypes of memory CD8+ T cells do shift in the absence of DRP1. Within the inguinal lymph node, OVA_257–264-specific memory CD8+ T cells of dLckCre × Drp1^flfl mice are predominately effector memory cells as defined by CD62L^loCCR7^lo, while the CD62L^hiCCR7^hi central memory cells are significantly reduced (Fig. 5I–L). Meanwhile, the fraction of OVA_257–264-specific memory CD8+ T cells resident to the lung is equivalent in dLckCre × Drp1^flfl and control mice (Fig. 5Q, R). Within the population of CD8+ T cells that are not circulating, tissue resident memory CD8+ T cells of the lung can be further identified by the expression of CD69 and CD103.²⁹ In the absence of DRP1, the highly retained CD69+ CD103+ double positive tissue resident memory T cells are equivalent between dLckCre × Drp1^flfl and control mice (Fig. 5S–U).

Fig. 5.

dLckCre × Drp1^flfl and Drp1^flfl mice develop equivalent resting memory CD8+ T cell populations after priming. Spleens (A–D), inguinal lymph nodes (E–L), and lungs (M–U) were collected from mice 28 d postpriming with αCD40, polyI:C, and OVA protein. Representative dot plots of spleen samples from Drp1^flfl (A) and dLckCre × Drp1^flfl (B) mice with percentage (C) and number (D) of CD44+ OVA_257–264 multimer + CD8+ T cells. Dot plots of inguinal lymph node samples from Drp1^flfl (E) and dLckCre × Drp1^flfl (F) mice with summary of percentage (G) and number (H) of CD44+ OVA_257–264 multimer + CD8+ T cells. Dot plots of inguinal lymph node samples from Drp1^flfl (I) and dLckCre × Drp1^flfl (J) mice with summary of percentage of CD62L–CCR7– effector memory (K) and CD62L+ CCR7+ central memory (L) OVA_257–264 multimer + CD8+ T cells. Dot plots of lung samples from Drp1^flfl (M) and dLckCre × Drp1^flfl (N) mice with summary of percentage (O) and number (P) of CD44+ OVA_257–264 multimer + CD8+ T cells. Histograms from Drp1^flfl and dLckCre × Drp1^flfl mice (Q) with percentage (R) of OVA_257–264 multimer + CD8+ T cells that were negative for intravenous CD45.2 labeling. Tissue resident memory cells in the CD45.2 IV–negative population of OVA_257–264 multimer + CD8+ T cells in the lung subset by CD69 and CD103 as dot plots from Drp1^flfl (S) and dLckCre × Drp1^flfl (T) mice with summary percentage of CD69+ CD103+ cells (U). (V) Number of CD44+ OVA_257–264 multimer + CD8+ T cells identified in recipient mice 5 d post challenge with αCD40, polyI:C, and OVA protein after adoptive transfer of memory CD8+ T cells. Individual values and mean ± SEM of n = 10 mice per group shown from 1 of 2 similar experiments. Statistics calculated using unpaired t tests. *P ≤ 0.05; ***P ≤ 0.001; ****P ≤ 0.0001.

We further questioned the functionality of DRP1-deficient memory CD8+ T cells by transferring an equivalent number of quiescent memory cells from dLckCre × Drp1^flfl into naïve recipient mice or from control Drp1^flfl mice into naïve recipient mice. These mice, containing equivalent DRP1-deficient or DRP1-replete memory T cells were then challenged with αCD40, polyI:C, and OVA protein. 5 d after challenge, the secondary OVA_257–264-specific CD8+ T cells resulting from dLckCre × Drp1^flfl cells were reduced by half compared with littermate control mice (Fig. 5V). Thus, in contrast to the complete secondary response seen in dLckCre × Drp1^flfl mice (Fig. 4D), the secondary response from adoptively transferred memory cells replicates the magnitude of the defect found in DRP1-deficient primary responses. These data suggest that in the absence of DRP1, secondary CD8+ T cells are impaired when removed from the DRP1-deficient environment. Further, this impairment does not appear to be related to mitochondrial health, as DRP1-deficient memory CD8+ T cells do not have altered mitochondrial mass or mitochondrial membrane potential compared with DRP1-replete memory CD8+ T cells (Supplementary Fig. 4E–H).

3.6. DRP1-deficient T cells cannot compete with WT T cells during primary, memory, or secondary responses

While the dLckCre × Drp1^flfl model provides an assessment of the role for DRP1 in T cells in general, DRP1 is deleted in both CD8+ and CD4+ T cells. Additionally, we noted a deeper reduction in the CD8+ T cell response in E8iERT2Cre × DRP1^flfl CD8+ T cell numbers than observed in dLckCre mice, suggesting a qualitative difference in CD8+ T cell responses when DRP1-deficient and DRP1-replete CD8+ T cells are present in the same host. To examine this further, we developed mixed bone marrow chimera mice in order to further address the CD8+ T cell intrinsic role of DRP1, as well as to evaluate whether the absence of DRP1 in CD8+ T cells results in a competitive disadvantage. In mixed bone marrow chimeras, 50% of the CD4+ T cells will be replete with DRP1 and will be able provide classical helper T cell functions to DRP1-deficient and replete CD8+ T cells. Cells resulting from the WT donor are CD45.2+ Thy1.1+, while the dLckCre × DRP1^flfl cells are CD45.2+ Thy1.2+. The resulting mixed bone marrow chimera mice were not exactly a 50/50 ratio of dLckCre × Drp1^flfl to Drp1^flfl (Fig. 6A); therefore, data from future time points were normalized to the reconstituted distribution, and raw data are included in Supplementary Fig. 6. Chimeras were primed as previously, rested to develop memory, and challenged using Adeno-OVA. As seen in dLckCre × Drp1^flfl mice (Fig. 1D), DRP1-deficient CD8+ T cells in mixed bone marrow chimera mice also demonstrate a reduced primary response (Fig. 6B) compared with their DRP1-replete counterparts, further supporting the notion that the deficit that arises from the absence of DRP1 is CD8+ T cell intrinsic. Interestingly, the control CD8+ T cells make a 6-fold larger primary response than DRP1-deficient T cells (Fig. 6B, E), a more extreme difference than observed in dLckCre × Drp1^flfl mice (Fig. 1D) and more equivalent to the reduction in response observed in E8iERT2CREx Drp1^flfl mice (Fig. 2D). Despite the significant impact to the primary response, mixed bone marrow chimera mice rested to memory ultimately develop equivalent numbers of OVA_257–264 memory CD8+ T cells from WT and dLckCre × Drp1^flfl donors (Fig. 6C, E), consistent with the relatively superior memory CD8+ T cell development in the absence of DRP1 observed in the dLckCre × Drp1^flfl mice. In contrast to the results from dLckCre × Drp1^flfl mice, the secondary response of mixed bone marrow chimeras was also dominated by DRP1-replete CD8+ T cells. The control response was approximately 2.5-fold greater than the DRP1-deficient CD8+ T cells (Fig. 6D, E). Thus, despite dLckCre × Drp1^flfl mice forming strong secondary responses, these cells expand weakly in the presence of WT secondary CD8+ T cells (Fig. 6F).

Fig. 6.

dLckCre × Drp1^flfl CD8+ effector T cells are diminished in the presence of WT effector CD8+ T cells. Mixed bone marrow chimera mice were primed with αCD40, polyI:C, and OVA protein and secondary responses initiated 28 d later by challenge with Adeno-OVA virus. (A) Relative chimerism within the naïve CD8+ T cell population from either the WT or dLckCre × Drp1^flfl donors was identified in the mice prior to priming. The magnitude of the ensuing CD44+ OVA_257–264 multimer + CD8+ T cell populations was determined by major histocompatibility complex-multimer staining of blood drawn day 7 postpriming (B), day 28 postpriming (C) or of splenocytes day 5 after secondary challenge (D). CD44+ OVA_257–264 multimer + CD8+ T cells originating from WT or dLckCre × Drp1^flfl donors as a ratio (E) or percentage within a specified time point (F). Individual values and mean ± SEM of n = 5 mice per group shown from 1 of 2 similar experiments. Statistics calculated using paired t tests or 1-way analysis of variance. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001. GMFI = Geometric mean fluorescence intensity; ns = not significant.

3.7. Contracting CD8+ T cells are not less prone to cell death or altered differentiation

The equivalency in the number of memory CD8+ T cells in control and DRP1-deficient mice, in multiple model systems, despite clear disparities in primary responses, suggests a difference in the efficiency of generating memory CD8+ T cells. For dLckCre × Drp1^flfl memory CD8+ T cells to recover from the impaired primary response, the effector CD8+ T cells likely have improved survival or altered differentiation toward memory development. As elongated mitochondrial phenotypes are utilized to promote survival during stress conditions,³⁰ we first hypothesized that the deletion of DRP1 allowed CD8+ T cells to better survive the programmed contraction phase. To assess if dLckCre × Drp1^flfl mice have reduced apoptosis, we stained cells for Annexin V 10 d postpriming, once the contraction phase had begun. We found no evidence for decreased phosphatidylserine exposure or apoptosis in live DRP1-deficient OVA_257–264-specific CD8+ T cells (Fig. 7A–D). The poly-caspase FLICA reagent also revealed no decrease in cell death of DRP1-deficient OVA_257–264-specific CD8+ T cells (Fig. 7E–G). Further, neither the fraction of OVA_257–264 T cells expressing the antiapoptotic molecule BCL2 nor the amount of BCL2 per cell, (Fig. 7H–J) was increased. These data indicate that neither a decrease in death nor an increase in survival during contraction 10 d postpriming explain the recovery of CD8+ T cells from dLckCre × Drp1^flfl mice after the primary response.

Fig. 7.

Contracting dLckCre × Drp1^flfl CD8+ T cells are resistant to cell death. Spleens were collected for flow cytometry from mice 10 d after priming with αCD40, polyI:C, and OVA protein. Results were pregated on single cells, lymphocyte scatter, viable cells, CD45+, CD8+, CD44+, and OVA_257–264 multimer+. Dot plots of spleen samples from Drp1^flfl (A) and dLckCre × Drp1^flfl (B) mice with percentage (C) and fluorescence intensity (D) of Annexin V + CD44+ OVA_257–264 multimer + CD8+ T cells. Histograms of spleen samples from Drp1^flfl and dLckCre × Drp1^flfl mice (E) with percentage (F) and fluorescence intensity (G) of poly caspase FLICA + CD44+ OVA_257–264 multimer + CD8+ T cells. Histograms of spleen samples from Drp1^flfl and dLckCre × Drp1^flfl mice (H) with percentage (I) and fluorescence intensity (J) of BCL2+ CD44+ OVA_257–264 multimer + CD8+ T cells. Individual values and mean ± SEM of n = 5 mice per group. Statistics calculated using unpaired t tests. *P ≤ 0.05. GMFI = Geometric mean fluorescence intensity.

Given the absence of evidence of differential survival, we hypothesized that a greater proportion of CD8+ T cells lacking DRP1 were destined to become memory precursors (CD127^hiKLRG1^lo;MPEC) over short-lived effector (CD127^loKLRG1^hi;SLEC) T cells. However, we found no significant differential SLEC:MPEC representation in the primary response between dLckCre × Drp1^flfl mice and control mice (Fig. 8A, B, and D). These data are further supported by comparable levels of TBET (Fig. 8E–G) and EOMES (Fig. 8E, F, and H) at 10 d postpriming, whereas increases would indicate preferential effector or memory differentiation, respectively.^31,32 Furthermore, there is no evident decrease in the terminal effector–associated transcription factor BLIMP1 (Fig. 8I–K) or increase in the memory cell–associated transcription factor BCL6 (Fig. 8I, J, and L).^33,34 Additionally, the control and experimental groups of OVA_257–264-specific CD8+ T cells have comparable expression of TCF1 (Fig. 8M, N), which is critical for self-renewing T cells.³⁵ Taken together, these indicators of T cell differentiation do not describe a method by which the absence of DRP1 promotes memory differentiation or supports short-lived effector differentiation.

Fig. 8.

dLckCre × Drp1^flfl CD8+ T cells do not preferentially differentiate to memory precursors. Spleens were collected from mice 7 d (A–D) or 10 d (E–N) postpriming with αCD40, polyI:C, and OVA protein. Dot plots of spleen samples from Drp1^flfl (A) and dLckCre × Drp1^flfl (B) mice with percentage of KLRG1+ (C) and CD127+ (D) CD44+ OVA_257–264 multimer + CD8+ T cells. Dot plots of spleen samples from Drp1^flfl (E) and dLckCre × Drp1^flfl (F) mice with percentage of TBET + (G) and EOMES + (H) CD44+ OVA_257–264 multimer + CD8+ T cells. Dot plots of spleen samples from Drp1^flfl (I) and dLckCre × Drp1^flfl (J) mice with percentage of BLIMP1+ (K) and BCL6+ (L) CD44+ OVA_257–264 multimer + CD8+ T cells. Histograms of spleen samples from Drp1^flfl and dLckCre × Drp1^flfl mice (M) with percentage (N) of TCF1+ CD44+ OVA_257–264 multimer + CD8+ T cells. Individual values and mean ± SEM of n = 5 mice per group. Statistics calculated using unpaired t tests.

4. Discussion

In the current study, we have extended previous investigations of the role of DRP1 in CD8+ T cell expansion and differentiation. We tested the hypothesis that restricting DRP1 in CD8+ T cells, promoting mitochondrial elongation and potentially redirecting their metabolic activity to one that favors OXPHOS, might provide an avenue for promoting memory CD8+ T cell development. In the absence of DRP1, the primary effector CD8+ T cell response is predictably limited, yet still surprisingly robust considering the role that DRP1 plays in mitochondrial segregation. Notably, the memory and secondary effector CD8+ T cell responses are not compromised, unless competition with DRP1-replete CD8+ T cells occurs. We found no evidence for alterations in cell death during primary effector CD8+ T cell contraction nor a shift toward memory precursor differentiation to explain these observations. While DRP1 deletion did increase elongation of mitochondria as expected, it did not enhance mitochondrial mass or oxygen consumption by activated CD8+ T cells. The strongest association for improved memory development in the absence of DRP1 is a reduction in the expression of the TCR and TCR signaling in naïve CD8+ T cells.

We anticipated that DRP1 would be important for the development of primary effector CD8+ T cells as DRP1 is involved during the processes of cellular division and T cell activation.^15,16 To date, the role of DRP1 in supporting T cell expansion has primarily been suggested by either in vitro stimulations or in short-term, tumor antigen–driven in vivo accumulation. We found a similar reduction in the number of OVA_257–264-specific CD8+ T cells 7 d after in vivo immunization, and importantly, novel to the studies presented here, further determined this to be intrinsic to CD8+ T cells, rather than an indirect consequence of a compromised CD4+ T cell response. We observed proliferative differences in DRP1-replete and deficient CD8+ T cells that could be attributed to differences in TCR expression and/or TCR signaling. While the majority of dLckCre × Drp1^flfl CD8+ T cells entered into proliferation after TCR stimulation, a small fraction of cells did not proliferate. During TCR engagement, DRP1 is needed to control the positioning of mitochondria at the immune synapse. Without the proximity of mitochondria to the immune synapse, the strength of TCR signaling is reduced,¹⁵ possibly accounting for both the reduced proliferative rate and the population of T cells that fail to expand after TCR engagement. We also noted a modest reduction in the steady-state levels of TCR on DRP1-deficient naïve T cells. While glycolytic activity has been shown to influence TCR recycling after stimulation,³⁶ the basis for reduced TCR expression without engagement is unclear but could be related to the role of DRP1 to actin remodeling. We hypothesized that the residual CD8+ T cells compensated for the loss of mitochondrial fission by producing new mitochondria. Mitochondria can be constructed through the process of mitochondrial biogenesis controlled by the master regulator PGC1α³⁷; however, we did not identify an increase in PGC1α within 2 d of in vitro stimulation.

A primary purpose of activated effector T cells is to proliferate and produce cytokines.³⁸ We evaluated how the absence of DRP1 would impact effector activities of CD8+ T cells. On the one hand, DRP1 deficient primary effectors produced less cytokines when stimulated via the TCR with OVA_257–264-pulsed target cells, suggesting that DRP1 might be needed for either the differentiation or activation of effector CD8+ T cells. However, on the other hand, bypassing TCR signaling with mitogen activation revealed that CD8+ T cells can activate a full suite of effector activities. Thus, we suspect that the absence of DRP1 does not prevent effector T cell differentiation, but rather likely limits effector activity via a reduced efficiency in TCR signaling, as discussed previously. Interestingly, secondary CD8+ T cells in dLckCre × Drp1^flfl mice are fully functional and capable of cytokine production without DRP1. Memory CD8+ T cells have a lower activation threshold than naïve CD8+ T cells³⁹ DRP1-deficient CD8+ T cells may be able to reach this threshold without the organization of mitochondria at the immune synapse by DRP1.

Rapid proliferation of memory cells is crucial to a timely recall response to secondary challenge.⁴ In addition to proliferation kinetics, memory T cells can rapidly respond to a secondary challenge due to their presence in peripheral tissues.²⁸ Interestingly, the proliferative impairment seen during primary effector development did not impede the magnitude of the secondary CD8+ T cell expansion in dLckCre × Drp1^flfl mice. Additionally, DRP1 was not needed for CD8+ memory cells to populate peripheral tissues such as lungs or secondary lymphoid organs including spleen and inguinal lymph nodes, despite previous observations indicating the absence of DRP1 limits the trafficking of effector T cells.¹⁸

In contrast to primary effectors, memory CD8+ T cells utilize an enhanced mitochondrial network in order to rapidly respond to a secondary pathogen challenge.¹¹ In the absence of DRP1, we found a comparable memory CD8+ T cell population forms, in contrast to the compromised number of primary effector T cells. In both dLckCre × Drp1^flfl mice and mixed bone marrow chimeras developed from dLckCre × Drp1^flfl cells, we found a high ratio of control cells over dLckCre × Drp1^flfl CD8+ cells during the primary response that returned to an equivalent ratio by the time quiescent memory establishes. At 10 d postpriming, during the contraction period, we were not able to identify decreased cellular death of dLckCre × Drp1^flfl CD8+ T cells as an explanation for the memory population that does develop. Further, the classical surface markers of CD8+ T cell effector or memory differentiation were not altered at 7 d postpriming, nor was the classical transcriptional programming associated with effector or memory differentiation altered at 10 d postpriming. Surprisingly, DRP1 deficiency does not provide additional SRC and also did not increase mitochondrial mass or membrane potential, which are thought to contribute to the metabolic advantages observed in memory T cells.^10,11 Mitochondrial elongation through pharmacological methods has been shown to also increase mitochondrial mass and OCR.¹⁰ DRP1 deletion also leads to elongated mitochondria and augments memory cell development but does not have the equivalent impact on OCR, suggesting that mitochondrial elongation is sufficient for the enhanced expansion found with memory T cells. Alternatively, the ability of mice with DRP1-deficient CD8+ T cells to develop a relatively elevated memory population might be related to the decreased TCR signaling seen in the absence of DRP1, as previous studies have clearly shown that the strength of TCR signaling can be a strong determinant in T cell differentiation and cytokine production capabilities.^40–43

In contrast to the results seen in dLckCre × Drp1^flfl mice, in which all CD8+ T cells lose DRP1 expression, in a competitive environment dLckCre × Drp1^flfl CD8+ T cells are not able to compete with DRP1-replete CD8+ T cells. The primary response of DRP1-deficient CD8+ T cells in mixed bone marrow chimeras is 6-fold lower, while the secondary response is 2.5-fold lower compared with the DRP1-expressing CD8+ T cells. While the basis for this observation remains to be mechanistically determined, in the mixed bone marrow chimera model, DRP1-deficient CD8+ T cells are potentially unable to compete for the uptake of cellular fuel. As memory cells primarily utilize fatty acid oxidation, DRP1 may be needed for optimal metabolism or uptake of fatty acids.¹⁸ Alternatively, in the absence of DRP1-mediated organization of mitochondria or structural proteins to the immune synapse, the CD8+ T cell interaction with antigen-presenting cells may not be as stable as interactions of WT cells. This could result in the diminished primary and secondary responses of DRP1 CD8+ T cells identified in the competitive setting.

Translationally, targeting DRP1 has the potential to improve memory T cell populations but needs to be evaluated in the context of the diminished primary response and the influence of competition. In particular, adoptive cell therapies or chimeric antigen receptor T cells have the most promising applications for targeting DRP1 in order to increase the number of persistent cells, but the timing and duration of DRP1 inhibition need to be carefully considered. Thus, pharmacological or inducible inhibition may provide the most tangible benefits. These applications allow for the temporal and cell-specific control of DRP1 inhibition during the ex vivo preparation of therapeutic cells, which would be essential to avoid the negative effects identified during the primary effector response. As deletion of DRP1 in CD8+ T cells affects many processes such as mitochondrial membrane maintenance and cell division,^16,17,19 transient inhibition of Drp1 could be utilized for the generation of persistent cells.

Author contributions

All authors conceived the project, designed experiments, interpreted data, and wrote the manuscript. M.G.S. performed experiments and analyzed data. F.M.M. performed and processed all microscopy imaging. T.N.J.B. supervised and acquired funding.

Supplementary material

Supplementary materials are available at Journal of Leukocyte Biology online.

Funding

This work was funded by National Institutes of Health grant R01CA166458 to T.N.J.B. M.G.S. was supported by National Institutes of Health training grant T32AI007496.