Larval T Cells Are Functionally Distinct from Adult T Cells in Xenopus laevis

In the amphibian Xenopus laevis, T cell differentiation in tadpoles and adult frogs is separated in time and occurs in different environments, which likely results in functionally distinct T cells (1, 2). For example, premetamorphic tadpoles in X. laevis are more susceptible than their adult counterparts to viral infection (e.g., ranavirus frog virus 3 [FV3]), which may in part be due to hyporesponsiveness of tadpole T cells (3). Similarly, Mycobacterium marinum (Mm) pathogens induce a tolerogenic immune response in tadpoles, contrasting with a robust T cell–mediated inflammatory response in adults (4).

Innate-like T (iT) cells represent a prominent component of the mammalian neonate immune system because, unlike conventional T cells, iT cells exhibit immediate effector function (5, 6). During early mammalian development, iT cells are represented by a variety of populations, including γδ T cell subtypes, MHC class I–like restricted mucosal-associated invariant T and invariant NKT cells, as well as innate-like CD4 and CD8 T cells known as virtual memory T cells (7, 8). As recently reviewed (9), tadpole T cells might represent a distinct population from their adult counterparts. In tadpoles, first peripheral T cells are detected at 8 postconception weeks and at day 15 after hatching (stage 52), respectively (reviewed in 2, 10). In the thymus of tadpoles, TdT expression is low in comparison with adults (11, 12). Accordingly, TCR diversity (including the CDR3 length) may be constrained in the tadpole. Indeed, the tadpole’s TCR α repertoire of CD8^neg and CD8^dim T cells is dominated by six invariant chains, implying a predominance of iT cell populations at this developmental stage (13), which is consistent with the suboptimal expression of polymorphic classical MHC class I protein until the onset of metamorphosis and the expansion of nonpolymorphic MHC genes expressed in tadpoles (14, 15). Two of these invariant TCR α-chains defining two distinct iT cell subsets have nonredundant roles in the tadpole’s resistance against ranavirus FV3 or Mm infection (13, 16).

In mammals, T cell activation and response magnitude result from the integration of signals coming from the TCR activation, APC-derived costimulatory signals, coinhibitory signals, and cytokines from the microenvironment (17–20). The TCR signaling strength can set the magnitude of the T cell responses as characterized by cytokine release and proliferation (17, 20). Ag-mediated TCR activation triggers a complex cytoplasmic signaling pathway involving Ca²⁺-calcineurin-NFAT, ERK1/2-AP-1, PKC-NF-κB, and AKT-mammalian target of rapamycin activation (reviewed in 17, 21). These TCR signaling components are evolutionarily conserved in jawed vertebrates (22–27).

Unlike the mammalian fetus, tadpoles are free living and devoid of maternal protection. The present work aimed to explore in X. laevis the mechanisms governing early developed T cells during immune responses against viral and mycobacterial pathogens.

All animals were obtained from the X. laevis Research Resource for Immunology at the University of Rochester (https://www.urmc.rochester.edu/microbiology-immunology/research/xenopus-laevis.aspx). All animal experiments were carefully handled with the prior approval and under the University of Rochester Committee on Animal Resources regulations (approval number 100577/2003-151).

Baby hamster kidney cells (BHK-21; American Type Culture Collection, CCL-10) were maintained in DMEM (Invitrogen) containing 10% FBS (Invitrogen), streptomycin (100 μg/ml), and penicillin (100 U/ml) in a 5% CO₂ atmosphere at 37°C, then 30°C for infection. FV3 was grown using a single passage through BHK-21 cells and purified by ultracentrifugation on a 30% sucrose cushion. Premetamorphic tadpoles (3 wk old, stages 54–55) were infected by i.p. injection of 6 × 10⁴ CFU Mm (stock no. PM2690 [4]) and 1 × 10⁴ PFU FV3 in 10 µl amphibian PBS (APBS), whereas adult frogs were infected by i.p. injection of 1 × 10⁶ CFU Mm or 2 × 10⁶ PFU in 100 µl APBS.

Uninfected control animals were mock infected with an equivalent volume of amphibian APBS. Tadpoles and frogs were injected with 5-ethynyl-2′-deoxyuridine (EdU; catalog no. A10044, Invitrogen, Waltham, MA) via i.p. injection ∼18 h prior to tissue collection for flow cytometry. Tadpoles were injected with 5 μl 1 μg EdU, and adult frogs were injected with 2.5 µg EdU per 1 g body weight. At the indicated days postinfection (dpi), animals were euthanized using 0.1 g/L tricaine methane sulfonate buffered with bicarbonate prior to dissection for immune cell extraction and flow cytometry as well as pathogen detection from the kidney and liver for FV3 and Mm infection, respectively, as previously described (16, 28).

Flow cytometry was conducted with freshly isolated tadpole and adult frog organs. For tadpoles, two same-tank replicates were pooled for each biological replicate, whereas each adult frog served as a biological replicate. To isolate splenocytes and thymocytes, the whole spleen and thymus were disrupted using a 100-μm pore size mesh and rinsed with APBS containing 1% BSA and 0.05% sodium azide, then labeled with mAbs. All mAbs were produced in-house by the University of Rochester Medical Center X. laevis Research Resource for Immunology. T cells were stained using the X. laevis pan-T cell markers CD5 and CD8 (29, 30), followed by secondary Abs (goat anti-mouse FITC (catalog no. A32723) and PE-conjugated streptavidin (catalog no. 12-4317-87), both from Thermo Fisher Scientific, Waltham, MA) at 1:300 and 1:100 dilutions, respectively. Cell viability was determined using 1:100 dilution of Ghost 510 Violet (catalog no. 13-08070-T100). Subsequently, cells were fixed with 2% paraformaldehyde and permeabilized with 0.2% Triton X-100 (catalog no. 0694, VWR, Radnor, PA), then incubated with a reaction mix consisting of 1 µl Cy5 Sulfo-Azide Dye (catalog no. A3330, Lumiprobe, Hunt Valley, MD), 50 µl sodium ascorbate, 10 µl copper sulfate, and 440 µl TBS (100 mM, 7.6 pH). Following intracellular EdU staining, each sample was washed twice and analyzed using an LSR II flow cytometer (BD Biosciences, Franklin Lakes, NJ). Fifty thousand events per sample were collected, and gating controls included unstained samples, secondary Ab–only controls, and single-stained controls. Flow cytometry data were analyzed using FlowJo version 10.8.2 (BD Biosciences) using the gating strategy shown in Supplemental Fig. 1A.

Splenocytes were harvested from adult frogs or tadpoles (3 wk old, stages 54–55), counted, and plated in Amphibian Serum Free Media (145 ml MSF medium, 75 ml water, 25 ml FBS, 5 ml penicillin-streptomycin, 100 µl kanamycin) at 100,000 cells per well in 96-well plates with a volume of 200 µl. Cells were treated with 0.5 µg/ml PHA (catalog no. 10576915, Thermo Fisher Scientific) dissolved in APBS or APBS control. After 2 h, 6 h, and 24 h, cells were harvested for RT-quantitative PCR (qPCR) analysis. After 48-h incubation, cells were treated with 40 µM EdU. At 72 h after PHA stimulation, cells were harvested for flow cytometric staining.

After 2 h, 6 h, and 24 h, splenocytes were harvested, and total RNA was extracted from the tadpole’s tissues using TRIzol reagent, following the manufacturer’s protocol (Invitrogen, Waltham, MA). For each sample, cDNA was synthesized using Moloney murine leukemia virus reverse transcriptase (Invitrogen) with oligo(dT) primers (Invitrogen). For RT-PCR, diluted cDNA was used to determine the expression levels of genes of interest by ΔΔCT value using an ABI 7300 Real-Time PCR System and PerfeCTa SYBR Green FastMix (Thermo Fisher Scientific) following the manufacturer’s protocol. Primers used are listed in Supplemental Fig. 1C. Relative gene expression levels were assessed using the ΔΔCT method as previously described (28). Briefly, expression levels were normalized to an endogenous housekeeping gene, gapdh.S (or ef1a.L when specified), then further normalized against the highest ΔCT among control and infected groups. The Gapdh gene is a useful housekeeping gene because of its consistent and high expression levels during viral and mycobacterial infection as well as ontogenesis (4). All the primers were validated prior to use by analysis of the size of the PCR product on a 2% agarose gel and qPCR melting curves (31).

For all experiments, statistical analyses were performed in GraphPad Prism 7 according to the experimental design of each experiment. A nonparametric Kruskal–Wallis test followed by a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli was used when analyzing time- and developmental stage–dependent effects of Mm and FV3 infection. The effects of PHA on the relative gene expression were analyzed using a Wilcoxon matched-pairs signed-rank test for each time point in tadpoles and frogs. A nonparametric Mann–Whitney U test was used to compare gene expression between tadpoles and adults at each time point.

To assess T cell proliferative response in vivo, we first used infection with the well-characterized ranavirus FV3. We measured the overnight EdU incorporation in T cells following FV3 i.p. inoculation. The time point in adult frogs was chosen according to previously published BrdU incorporation in CD8⁺ T cells during FV3 infection (32). In tadpoles, a longer time point (9 dpi) was added because the immune response during FV3 infection is generally delayed in tadpoles as it is in mouse neonates compared with adults (3, 33, 34). As expected, FV3 infection in adult frogs led to a significant increase in frequency of proliferating splenic EdU⁺ T cells, both CD8⁺ and CD8⁻ CD5⁺ T cells at 3 dpi and sharply increasing at 6 dpi (Fig. 1A). On average, 10% of T cells were EdU⁺ at 6 dpi, which corresponds to what was previously observed using BrdU (Fig. 1C). In contrast, no to very limited increases in the proportion of EdU⁺ CD8⁺ and CD8⁻ CD5⁺ T cells were induced during FV3 infection in tadpoles even at 9 dpi (Fig. 1B–1D). This was not due to some general defect or inefficient EdU incorporation in tadpoles, because comparable levels of proliferating EdU⁺ thymocytes were detected between tadpoles and adult frogs (Fig. 1A, 1B, last panel). FV3 infection in tadpoles was confirmed by quantifying the FV3 genome copy number (Fig. 1E).

To determine whether the poor T cell proliferative response detected in tadpoles was limited to this viral pathogen, we extended our study to a bacterial pathogen, Mm. On the basis of our previous study showing an increase in the CD8⁺ T cell numbers in the X. laevis adult spleen at 6 and 12 dpi (4), we determined the overnight EdU incorporation following Mm inoculation at these two time points. Although there was some increase in the frequency of EdU⁺ T cells in the adult spleen at 6 dpi that did not reach statistical significance, a marked proliferative response slightly higher for CD8⁺/CD5⁺ T cells (12%, on average) than CD8⁻/CD5⁺ T cells (8%, on average) was detected at 12 dpi (Fig. 2A, 2C). Again, as with FV3 infection, very limited T cell proliferative activity was detected under steady state in tadpoles at both 6 and 12 dpi. It is notable that for uninfected control animals, the proportion of EdU⁺ CD8⁻ CD5⁺ T cells was significantly higher in tadpoles than in adults (Figs. 1 and 2).

To investigate whether the distinctive peripheral T cell proliferative capacity between tadpoles and adult frogs was intrinsic, we conducted in vitro proliferative assays using PHA stimulation. It is notable that, to date, PHA stimulation assays in X. laevis have mostly been determined by thymidine incorporation on total splenocytes (35, 36). We stimulated splenocytes for 3 d adding EdU for the last 16 h. Interestingly, unlike the in vivo response to pathogens, both tadpole CD8⁺ and CD8⁻ CD5⁺ T cells showed a strong proliferative response to PHA stimulation with an average 40% EdU⁺ cells at 3 d after stimulation (Fig. 3A, 3C). In contrast, adult splenic T cells were less responsive with less than 10% T cells with EdU⁺ signal, on average. The proportion of EdU⁺ T cells was also significantly higher in unstimulated tadpole splenocyte cultures compared with adult cultures.

To further investigate possible intrinsic differences between tadpole and adult frog T cells, we analyzed the expression response of a set of key genes involved in T cell activation and TCR signaling. Tadpole and adult frog splenocytes were stimulated with PHA during 2 h, 6 h, and 24 h for gene expression analysis by RT-qPCR. Note that because X. laevis is allotetraploid, some genes are present and active on two sets of chromosomes distinguished by their length, long (L) and short (S), and, by convention, an “L” or “S” is appended to the corresponding genes (37, 38). Gapdh.S gene expression was used as a stable reference gene to compare stimulated and unstimulated tadpole and frog splenocytes at 2 h and 6 h (Supplemental Fig. 1B). Gapdh.S gene expression was also used to compare stimulated tadpoles and frog splenocytes at all time points. Due to unexpected variability of gapdh.S gene expression at 24 h, the ef1a.L gene was used as a reference gene at this time point (Supplemental Fig. 1B).

Several gene markers of TCR signaling and immediate-early genes known to be induced upon TCR activation were differentially expressed upon PHA stimulation in adult and tadpole splenocytes (Fig. 4). Notably, zap70.L encoding the ZAP ortholog that plays a critical role in T cell signaling was significantly upregulated with PHA at 6 h and 24 h in both tadpoles and adults. Among immediate early genes induced by TCR activation, egr2.L gene expression was significantly upregulated at all time points in tadpole and adult splenocytes. The kinetics of the Fos.L gene expression response to PHA stimulation differed between tadpole and adult splenocytes. Indeed, fos.L gene expression was significantly increased transiently only at 2 h in tadpoles, whereas in adults, it increased later at 6 h and remained higher at 24 h. Transcript levels of nur77.L were significantly induced by PHA stimulation, again slightly faster in adults than in tadpoles, and remained high at later time points. Expression of the Irf4.L gene, another transcription factor induced by TCR engagement, was also significantly upregulated in adults and tadpoles at all time points except 2 h in adults (p = 0.1094).

Owing to the importance of IL-2 in the T cell proliferative response, we assessed changes in the expression of genes encoding IL-2 receptors α, β, and γ as well as IL-2 (Fig. 5) (37). There was no real difference between tadpoles and adult frogs in expression response IL2 receptor–encoding genes. Only the transcript levels of il2ra.L were markedly higher in adult frogs than in tadpoles. However, il2.L transcripts decreased at an earlier time point in adult splenocytes (6 h) compared with tadpoles, which showed a decrease at 24 h.

We then thought to assess the expression of some relevant T cell effector genes. We first investigated the expression of the Il-4/13 gene, which, in Xenopus as in fish, shares homology to both Il-4 and Il-13 in avians and mammals, which likely arose by tandem duplication (39). X. laevis il-4/13.S gene expression was significantly upregulated after 6 h of stimulation in adults and tadpoles (Fig. 6). However, at 24 h, PHA significantly increased the relative expression of il4/13.S in adults only. Because expression of the Ifng gene ortholog ifng.S is delayed in tadpoles compared with adult frogs upon FV3 and Mm infection (16, 33), we determined its expression profile upon PHA stimulation. Consistent with these in vivo studies, transcript levels of ifng.S significantly increased more rapidly in adults (6 h) than in tadpoles (24 h).

Further comparison of the relative expression levels of PHA-stimulated splenocytes between tadpoles and adults at 6 h and 12 h revealed that the relative expression levels of nur77.L, irf4.L, il4/13.L, and ifng.S were all significantly higher in adult frog splenocytes (Fig. 7). Finally, as previously reported, we found that tadpole splenocytes were characterized by a lower frequency of CD5⁺ T cells and exhibited a lower cell surface CD5 staining intensity (Fig. 8) (30, 40).

In X. laevis, T cells produced in tadpoles appear functionally limited in terms of Ag recognition (i.e., TCR repertoire is limited in both gene usage and N-region diversity) (9). Such an assumption is based mainly on reverse genetic approaches and in vitro studies because in vivo assays to assess T cell responsiveness are limited in ectothermic vertebrates.

The present study shows that larval X. laevis T cells exhibit limited capacity to proliferate in vivo in the spleen (the only secondary lymphoid organ in Xenopus) during microbial infection. These data are consistent with previous findings that during Mm and FV3 infection, tadpoles develop a delayed and limited inflammatory response (Tnfα;, Ifnγ, and Il1β gene upregulation) and have limited capacity to restrain pathogen dissemination (4, 33, 41, 42). Evidence that tadpole T cells play a critical role in the host resistance against FV3 and Mm have been obtained by a reverse genetic approach (shRNA and CRISPR/Cas9) targeting the invariant Vα6-Jα1.43 (iVα6) restricted by nonclassical MHC1-UBA10.1.L for FV3 and Vα45-Jα1.14 TCR (iVα45) associated with MHC1-UBA4.L for Mm (4, 13, 16, 43, 44). Thus, it is possible that susceptibility to some pathogens in X. laevis tadpoles may in part be due to some hyporesponsiveness of their T cells in vivo. Whether this hyporesponsiveness is a result of a more limited TCR repertoire, which includes fewer virus- or bacteria-specific receptors than the adult repertoire, and/or some immaturity of larval APCs remains to be determined.

In vitro X. laevis tadpole T cells show higher proliferation capacity under PHA stimulation, as previously reported in X. laevis by thymidine incorporation assay (36). Similarly, in mammals, neonatal CD8 and CD4 T cells proliferate faster and sooner than their adult counterparts in response to TCR activation (45–48). PHA triggers T cell activation and proliferation by binding to cell surface glycoproteins, including the TCR and the costimulatory receptor CD2 (49–51). T cell proliferation in tadpoles and frogs is substantiated by the significant increase of relative expression at 6 and 24 h of the T cell–specific kinase zap70.L. Zap70 is a 70 kDa tyrosine kinase critically involved in the proximal TCR signaling (21). TCR signaling in X. laevis was inferred by measuring the relative expression of established immediate early genes (fos, egr2, irf4, and nur77) rapidly transcribed upon TCR activation. Early immediate genes are characterized by a rapid protein synthesis–independent induction of their transcription (52–57). We found that Nur77, Irf4, and Fos genes but not erg2 were differentially upregulated in X. laevis tadpole and adult splenocytes. Irf4 and Nur77 gene expression in murine and human T cells is proportional with the Ag affinity and, therefore, with the TCR signaling strength (58–60). Therefore, our data suggest that tadpole TCR signaling is hyporesponsive in terms of amplitude and duration compared with adult T cells. Although neonatal mice and human conventional naive T cells show an enhanced TCR signaling pathway and enhanced proliferative capacity in comparison with their adult counterparts (48, 61), hyporesponding TCR signaling has been described in the Ag-experienced memory T cells (58, 62), the early developing innate-like γδ T cells and αβ intraepithelial T cells (63) as well as in self-reactive naive T cells (64). Erg2 and Fos genes were also differentially upregulated in tadpole and adult splenocytes. Our results show that although fos.L expression is not upregulated at 24 h in tadpole T cells, erg2.L as well as irf4.L and nur77.L expression is still upregulated by PHA stimulation, indicating that TCR signaling pathways are regulated differently between tadpole and adult T cells. In mammals, Fos interacts with Jun to form AP-1 transcription factor complex, which is involved in the TCR signaling transduction (21, 65). These results suggest an early and late downregulation of the AP-1–dependent transcription after TCR activation in tadpole T cells only (21, 65). Similarly, AP-1–dependent transcription was found to be reduced in naive neonatal CD4 T cells (61). Although we cannot rule out that PHA binds to other cells, the measurement of EdU incorporation by flow cytometry indicates that PHA stimulates mostly X. laevis T cells as in mammals.

Our in vitro and in vivo results suggest that X. laevis tadpole T cells have higher homeostatic proliferation than their adult counterparts. In mammals, T cell homeostasis is dependent on self-antigen–MHC–TCR interaction and members of γc family (IL-15 and IL-7) receptors (66). Higher homeostatic proliferation characterizes memory T cells and early developing T cells, which are more self-reactive and which preferentially differentiate in innate-like memory cells such as virtual memory T cells (17, 47, 67, 68). Together, the hyporesponsive gene expression involved in TCR signaling in response to PHA stimulation, as well as the higher proliferation under homeostasis and under TCR activation, suggest that tadpole T cells have a memory-like phenotype, whereas adult frog T cells have a naive/conventional phenotype as in mammals. In mice, AP-1–dependent transcription, as well as nur77 (also known as Nr4a1) and irf4 are associated with T cell exhaustion and anergy (69). Likewise, it is possible that the lower proliferative capacity of X. laevis adult T cells in vitro following PHA stimulation results from a more naive phenotype, which, differently from innate-like or memory-like T cells, requires coactivator activation and sufficient IL-2 for optimal T cell response. Therefore, our data reinforce previous work indicating that tadpole T cells have prominent innate-like T cell features (13, 16). Innate T cells, memory T cells, and neonatal T cells are typically activated by innate or “stress” stimuli such as cytokines and NK and pathogen recognition receptor activation (8, 64, 70–76). Similarly to mammals, it is likely that such intrinsic differences between tadpole and adult T cells is related to ontogenesis and their different progenitor origins (9).

The authors have no financial conflicts of interest.

We thank Tina Martin for the expert animal husbandry.