Transcriptional Memory-Like Imprints and Enhanced Functional Activity in γδ T Cells Following Resolution of Malaria Infection

γδ T cells play an essential role in the immune response to many pathogens, including Plasmodium. However, long-lasting effects of infection on the γδ T cell population still remain inadequately understood. This study focused on assessing molecular and functional changes that persist in the γδ T cell population following resolution of malaria infection. We investigated transcriptional changes and memory-like functional capacity of malaria pre-exposed γδ T cells using a Plasmodium chabaudi infection model. We show that multiple genes associated with effector function (chemokines, cytokines and cytotoxicity) and antigen-presentation were upregulated in P. chabaudi-exposed γδ T cells compared to γδ T cells from naïve mice. This transcriptional profile was positively correlated with profiles observed in conventional memory CD8+ T cells and was accompanied by enhanced reactivation upon secondary encounter with Plasmodium-infected red blood cells in vitro. Collectively our data demonstrate that Plasmodium exposure result in “memory-like imprints” in the γδ T cell population and also promotes γδ T cells that can support antigen-presentation during subsequent infections.


INTRODUCTION
gd T cells are unconventional T cells that display characteristic features of both innate and adaptive immunity. Their capacity to respond rapidly to non-peptide antigens in an MHC-independent manner places them as part of the innate first line of defense against numerous pathogens. Additionally, emerging evidence supports the concept that gd T cells also display memory T cell-like abilities. This includes prolonged recall responses upon reinfection in various disease and vaccine models, which contribute to protective immunity (1)(2)(3)(4)(5)(6). Recent studies have now started to delineate a more in-depth understanding of these adaptive-like gd T cells. For example, it has been described that the TCR of tissue-resident gd T cells has an intrinsic ability to distinguish between distinct antigen-stimulus and in this way promote either clonal or non-clonal responses (7) whereas adaptive-like gd T cells found in peripheral human blood are suggested to be restricted to specific subsets of the gd T cell population (8).
Plasmodium infection, which is responsible for the induction of malaria in humans, elicits a multifaceted response activating a wide range of immune cells, including gd T cells. Extensive evidence shows that gd T cells are part of the immediate innate response during human malaria infection where they are found to be cytotoxically active and produce cytokines associated with both protective immunity and symptomatic episodes (9)(10)(11)(12)(13)(14)(15). The underlying mechanisms by which gd T cells either contribute to beneficial outcomes in the host or mediate pathogenesis remain to be fully elucidated.
In addition to human infections, gd T cells are also highly involved in the immune response to murine malaria. In mice, they are a major source of cytokines and contribute to parasite clearance (16)(17)(18)(19)(20)(21) and are essential for protective immunity following vaccination (22). This makes murine malaria infection models a useful platform to explore fundamental immunological questions related to immune populations, such as gd T cells, in an infectious disease setting. P. chabaudi infection in C57BL/6 mice is a selfresolving infection, and this infection model has been used to successfully elucidate various aspects of gd T cell biology. gd T cells proliferate extensively in response to P. chabaudi infection and mice lacking gd T cells experience exacerbated parasitemia (20,(23)(24)(25). More recently, gd T cells from chronically infected mice were described to produce inflammatory chemokines such as CCL3 and CCL5 and also importantly m-CSF, which was vital to the control of recrudescence (18) suggesting that "antigen-experienced" gd T cells play a role in the suppression of parasitemia in chronic infection. These studies further emphasize that gd T cells are readily activated during acute Plasmodium infection. However, the lasting effect that Plasmodium exposure has on these cells and how this shapes the gd T cell population is still inadequately understood. Consequently, we used the P. chabaudi murine malaria infection model to investigate transcriptional profiles of gd T cells from naïve and malaria-exposed mice, 12 weeks after completion of anti-malarial drug treatment. Our findings revealed that antigen-experienced gd T cells display a transcriptional profile that shares features with that of conventional memory CD8 + T cells and have enhanced functional capacity. Thus, our data support the notion that gd T cells differentiate and acquire a memory-like phenotype after infection. These observations advance our basic understanding of unconventional T cell biology and establish novel molecular qualities in these cells as a result of infection.

Mice and Mouse Infection
Female C57BL/6 mice aged 6-8 weeks were infected with 5 x 10 4 Plasmodium chabaudi iRBC intravenously. All mice (both infected and naïve mice) were drug-treated on day 14 p.i. or at an equivalent time for naïve mice with an intraperitoneal injection of chloroquine (CQ; 10 mg/kg) and pyrimethamine (10 mg/kg) followed by CQ (0.6 mg/ml) and pyrimethamine (70 µg/ml) containing water for 5 days. Spleens and livers were removed 12 weeks after completion of drug treatment. The experimental design is summarized in Figure 1A. Organs from drug-treated naïve mice were used as controls. All procedures involving mice were approved by the Walter and Eliza Hall Institute animal ethics committee (2015.020).

In Vitro Cell Stimulation
Single cell suspensions from spleen or liver were prepared as previously described (26). Wholeblood from P. chabaudiinfected donors were obtained during the dark cycle to obtain mature parasites (27). The blood was washed in RPMI and 0.5-1 ml of blood in medium was overlayed onto 12.17 ml of a 74% percoll gradient as described in (28) and centrifuged at 5000 g for 20 min at room temperature. IRBCs were collected from the interface and washed with culture medium. Isolated iRBCs were co-incubated with splenocytes and liver lymphocytes at a ratio of 1:1 for 24 h. Brefeldin A (Sigma, St. Louis, MO) and GolgiStop (BD Biosciences, San Jose, CA) were added for the final 8 h of incubation.

Adoptive Transfer
Single cell suspensions from spleens were prepared from naïve or P. chabaudi-exposed mice. gd T cells were isolated using TCRgd T cell isolation kit (Miltenyi Biotec, Australia) according to manufacturer's instructions. Isolated gd T cells were adoptively transferred (1x10 6 /mouse) into recipient C57BL/6 or RAG-1 mice, 1 day post-infection with 5 x 10 4 P. chabaudi iRBC intravenously. Parasitemia was measured daily by thin blood smears after Giemsa staining.

Library Preparation and Transcriptome Sequencing
EM gd T cells from five naïve control mice and five mice that had been previously infected with P. chabaudi and then drug-treated to clear the infections were FACS sorted. Total RNA was isolated from sorted cells using the Isolate II RNA mini kit (Bioline, London, UK) according to manufacturer's instructions. RNA was quantified using the Agilent TapeStation 2200 system (Santa Clara, CA). An input of 1 ng of total RNA were prepared and indexed separately for sequencing using the CloneTech SMART ultra-low RNA input Prep Kit (Illumina, San Diego, CA) as per manufacturer's instruction. The indexed libraries were pooled and diluted to 1.5pM for paired end sequencing (2 x 76 cycles) on a NextSeq 500 instrument using the v2 150 cycle High Output kit (Illumina) as per manufacturer's instructions. The base calling and quality scoring were determined using Real-Time Analysis on board software v2.4.6, while the FASTQ file generation and demultiplexing utilized bcl2fastq conversion software v2.15.0.4. Paired-end 75bp. Between 16 and 56 million read pairs were generated for each sample and reads were aligned to the Mus musculus genome (mm10) using the Subread aligner (29). The number of read pairs overlapping each mouse Entrez gene was summarized using featureCount (30) and Subread's built-in NCBI gene annotation. Genes were filtered using filterByExpr function in edgeR (31) software package. Genes without current annotation and Immunoglobulin genes were also filtered. Differential expression (DE) analysis was undertaken using the edgeR and limma (32) software packages. Library sizes were normalized using the trimmed mean of M-values (TMM) method (33). Log2 foldchanges were computed using voom (34). Differential expression was assessed relative to a fold change threshold of 1.5 using the TREAT (35) function, a robust empirical Bayes procedure (36) implemented in the limma package. The false discovery rate (FDR) was controlled below 0.05 using the method of Benjamini and Hochberg (37). Over-representation of Gene Ontology (GO) terms for the differentially expressed genes was identified using the goana function in limma package. Barcode plots illustrating the enrichment of interested pathway genes were drawn using the barcode plot function in limma package (38).
A B D C FIGURE 1 | Increased frequency of IFNg + CD107a + gd T cells in previously infected mice. (A) C57BL/6 mice were infected with P. chabaudi and then drug-treated with chloroquine and pyrimethamine 2 weeks later. Twelve weeks following completion of drug-treatment cells were isolated and stimulated with iRBCs or uRBCs and frequencies of IFNg + and/or CD107a + cells were assessed. (B) Representative flow cytometry plots illustrating the gating strategy. Frequencies of IFNg + and/or CD107a + (C) splenocytes, and (D) liver lymphocytes from previously infected mice (P. chabaudi black squares, n=14) and naïve control (white circles, n=10) after stimulation. In the pie chart the data are presented as the frequency of IFNg + CD107a + (blue), IFNg + CD107a -(red) and IFNg -CD107a + (green) gd T cells in each group following uRBC background subtraction. The data in the scatter plot are presented as mean ± SD following uRBC background subtraction. The data represent results from two independent experiments. Statistical analysis was performed using Student's t-tests. ***P < 0.001.

Kumarasingha et al.
Transcriptional Memory-Like Imprints in gd T Cells Frontiers in Immunology | www.frontiersin.org October 2020 | Volume 11 | Article 582358

Statistical Analysis
Statistical analyses were performed using Prism 8.0 (GraphPad software, San Diego, CA) Flow cytometry data was analyzed using the Student's t-test. Statistical significance was considered P ≤ 0.05.

RESULTS
Increased Frequencies of Multifunctional gd T Cells in Drug-Cured P. chabaudi-

Exposed Mice
The hallmark of memory T cells is increased functional capacity upon secondary encounter with specific antigen, which commonly includes IFNg production and cytotoxic activity. To establish whether similar responses were generated in gd T cells following Plasmodium infection, we compared responses of naïve and preexposed gd T cells upon antigen re-encounter. Since spleen and liver are central to the immune response to P. chabaudi infection (39,40) and are organs that have previously been shown to contain tissue resident innate memory cells (41,42), we assessed gd T cell responses in both of these organs. To that end, C57BL/6 mice were infected with P. chabaudi and drug-cured on day 14 postinfection (p.i.) to clear parasitemia completely. Twelve weeks after completion of drug-treatment spleens and livers were harvested ( Figure 1A). Splenocytes and liver lymphocytes were subsequently isolated and stimulated in vitro with P. chabaudi-infected red blood cells (iRBC) or uninfected RBC (uRBC) as background controls. Cells from naïve mice were included to measure baseline responses. After a 24 h incubation, CD107a surface expression (as a measure of cytotoxic activity) and IFNg production were assessed by flow cytometry ( Figure 1B). We found that a significantly higher frequency of gd T cells that were both CD107a + and produced IFNg were present in the spleens of previously infected mice compared to naïve mice ( Figure 1C, P< 0.0001). No significant differences were observed with gd T cells that produced only IFNg or were CD107a + . Similarly, no significant differences in functionality were detected in the liver-derived gd T cells from pre-exposed P. chabaudi-infected mice and naïve mice ( Figure 1D). This showed that P. chabaudi infection resulted in the induction of multifunctional memory-like gd T cells.

Responding gd T Cells Express an Effector Memory-Like Phenotype
Previous studies indicate that the gd T cells that provide effector functions during acute malaria infection express surface markers that resemble conventional ab T effector memory cells (18,43).
To assess the phenotype of the responding gd T cells of previously exposed mice after full resolution of infection, we stimulated spleen-derived gd T cells from drug-treated mice or naïve mice in vitro and stained the cells for the surface markers CD62L and CD44. This enabled the gd T cells to be subdivided into CD62L + CD44naïve cells, CD62L + CD44 + central memory cells (CM) and CD62L -CD44 + effector memory cells (EM; Figure  2A). The frequency of IFNg + CD107a + double positive gd T cells in each subset was assessed in both groups of mice.
Representative flow cytometry plots of these responses are presented in Figure 2B. Upon stimulation with iRBC, responding gd T cells were found to predominantly express an EM phenotype and frequencies of IFNg + CD107a + EM gd T cells were significantly higher in previously P. chabaudi-infected mice compared to naïve control mice (P< 0.0001; Figure 2C). This demonstrated that gd T memory-like responses were specifically confined within the EM subset. Furthermore, the increase in frequency of responding cells did not reflect an overall increase of EM gd T cells in the pre-exposed mice as assessment of the gd T cell composition showed no differences in frequencies ( Figure  2D) or cell numbers ( Figure 2E) of naïve, CM or EM gd T cells between P. chabaudi exposed mice and uninfected controls.
Transcriptional Profile Changes in EM gd T Cells From Drug-Treated P. chabaudi Exposed Mice Compared to EM gd T Cells From Naïve Mice We have shown that gd T cells expressing an EM-phenotype are reactivated upon re-encounter with P. chabaudi iRBC in previously infected mice ( Figure 2). As the frequency and number of EM gd T cells in the spleens were not different between the naïve control group and the pre-exposed mice, this indicated that this memorylike enhanced responsiveness was due to intrinsic changes of the cells. To investigate this, EM gd T cells were FACS-sorted from mice 12 weeks after they had been drug treated to clear P. chabaudi infection (n=5) and from naïve mice (n=5; Figure 3) and RNAsequencing was used to examine transcriptional profiles. A total of 207 differentially expressed (DE) genes in P. chabaudi pre-exposed EM gd T cells compared to EM gd T cells from naïve mice were observed relative to a fold change threshold of 1.5 (Supplemental Table 1). Expression levels and log-fold changes were plotted in a Mean-Difference (MD) plot ( Figure 3A) of which 96 genes were significantly upregulated (indicated in red) and 111 genes were significantly down regulated (indicated in blue). The upregulated genes included MHC class II-related genes (H2-Dmb2 and H2-A) and also IFNg and NKg7, which corresponded to the observed functional phenotype of enhanced IFNg production and cytotoxicity in the pre-exposed EM gd T cells ( Figure 2). The chemokine genes (CCL3, CCL4 and CCL5) were also upregulated in these memory-like gd T cells, which is similar to what had previously been reported to be upregulated in gd T cells during an active infection (Mamedov 2018). Cytokine receptor genes (Il1r and Il23r), scavenger-receptor gene (Cd163l1) and transcription factor gene (Sox13) were among the down regulated genes. The top 75 DE genes are summarized in a heatmap presenting up-and down regulated genes in each mouse ( Figure 3B). Collectively, this shows that malaria-infection causes significant transcriptional changes in the EM gd T cell population, which is still observed in absence of an active infection.

Genes Involved in Antigen Presentation
and Processing Are Upregulated in Pre-Exposed EM gd T Cells To understand the biological processes affected by previous exposure to malaria in the EM gd T cell population, gene ontology (GO) pathway analysis was performed. Among the 20 most highly enriched GO terms in the upregulated biological processes, seven were associated with antigen-processing and presentation. In addition, genes were enriched for processes involving positive regulation of acute inflammatory responses and response to IFNg Supplemental Figure 1). Barcode plots and bar plots illustrating the enrichment of all genes in selected pathways showed that antigen-processing and presentation pathway included upregulation of MHC class II-related genes (H2-Aa, H2-Dmb2, H2-Ab1, H2-Eb1, H2-Dmb1), genes that support antigenprocessing and presentation (Clec4a2, Flt3, Cd74, Ifng) and genes for FC receptor expression (Fcrgr2b, Fcer1g; Figure 4A). There

A B
D E C FIGURE 2 | In vitro re-stimulated and activated gd T cells express CD44, but lack CD62L expression. Splenocytes from previously infected and drug treated mice and naïve controls were restimulated in vitro with iRBC or uRBCs. Representative contour plot to (A) distinguish between CD62L + CD44 -(Naïve), CD62L + CD44 + (CM), and CD62L -CD44 + EM gd T cells. (B) Representative contour plots showing frequency of IFNg + CD107a + gd T cells for each subset after 24 h stimulation with either iRBC or uRBC from naïve or P. chabaudi pre-exposed mice. (C) Summary of IFNg + CD107a + naïve, CM, and EM gd T cells after iRBC stimulation following subtraction of background levels determined from uRBC stimulations in previously P. chabaudi-infected mice (filled squares; n=14) and naïve controls (open circles; n=10). Overall (D) frequency and (E) number of gd T cells per spleen of naïve, CM and EM gd T cells (mean±SD) in naïve or P. chabaudi pre-exposed mice. The data represent results from two independent experiments. Statistical analysis was performed using Student's t-tests ****P < 0.0001. were also enrichment of genes that suggested an increased responsiveness to IFNg stimulation as shown by upregulation of chemokine and cytokine genes (Ccl3, Ccl4, Ccl5, Ifng, Xcl1), interferon induced transmembrane protein genes (Ifitm2, Ifitm3) and MHC class II-related genes (H2-Aa, H2-Ab1, H2-Eb1), but down regulation of IL23r ( Figure 4B). In addition, gene enrichment analysis suggested that pre-exposed EM gd T cells have the potential to contribute to a sustained inflammatory response as shown by upregulation of Fcer1a, Alox5bp. Ptgs2, Fcer1g, and Ccl5 combined with down regulation of Adam8 ( Figure 4C). Some of the most significantly down regulated biological processes included cell-substrate adhesion and cellular response to stimulus (Supplemental Figure 1B). Considering that responsiveness to IFNg stimulation was increased (Supplemental Figure 1A), decrease in the biological process of cellular response to stimulus suggests that the pre-exposed EM gd T cell population is modulated to only respond to specific conditions such as presence of IFNg. Barcode plots and bar plots illustrating the enrichment of all genes in these down regulated pathways showed that a total of 85 DE genes were represented in the cellular response to stimulus ( Figure 5A). The three most down regulated genes in this pathway were Itgb4, Plxnd1, and Tspan2, which are all associated with signal transduction and cell-cell signaling. The most upregulated gene in this pathway was Fcer1a, which has been associated with an immune suppressive role in APCs (44). The enrichment of all genes in the cell-substrate adhesion pathway included down regulated integrin genes (Itgb4, Itga5, Itgb5), protein kinases (Trmp7, Slk) and genes associated with cell recruitment, adhesion and migration (Adam8, Jag1, Lamc1, L1cam) whereas Epdr and Smoc2 genes were upregulated ( Figure 5B).

Differentially Expressed Genes in Pre-Exposed EM gd T Cells Are Positively Correlated With Differentially Expressed Genes in Resting Memory CD8 + T Cells
A previous study demonstrated that conventional CD8 + memory T cells have distinct transcriptional profiles, even in a resting state (i.e. without re-stimulation) that significantly differ from those of their naïve counterparts (45). As we had also performed transcriptional analysis from pre-exposed but resting cells, we wanted to examine similarities between the two transcriptional profiles and we compared our DE expression data (Supplemental Table 1) with the previously described signature defining resting CD8 + memory T cells (45) (Russ et al. Supplemental Table 2). A total of 43 DE genes were represented in both gene sets, of which 32 were upregulated and 11 were down regulated DE genes ( Figure 6A). These overlapping genes presented in a heatmap ( Figure 6B) included genes that were associated with hallmark functions of conventional memory T cells such as cytokine/chemokine production and cytotoxicity (Ccl4, Ccl5, Ccl3, Ifng, Nkg7). Genes involved in antigen presentation and processing (Clec4a2, Fcgr2b, H2-Aa, H2-Dmb2, H2-Ab1, H2-Eb1, Cd74, H2-Dmb1, Fcer1g), which was a prominent transcriptional signature of the memory-like EM gd T cell DE gene set, also overlapped with the DE genes from CD8 + memory T cells. Furthermore, enrichment analysis A B FIGURE 3 | RNA-sequencing of EM gd T cells from P. chabaudi pre-exposed mice and naïve controls. EM gd T cells from drug-treated naïve mice (n=5 donors) and P. chabaudi pre-exposed mice (n=5) were FACS sorted followed by RNA extraction and RNA-sequencing. Differential gene expression for P. chabaudi over naïve mice was summarized in (A) mean-difference (MD) plot of log2 expression fold-changes against the average log-expressions for each gene. The differentially expressed (DE) genes relative to a fold change threshold of 1.5 are highlighted, with points colored in red and blue indicating up-and down regulated genes respectively. (B) Heatmap of the expressions of the top 75 DE genes between P. chabaudi and naïve mice. Each vertical column represents genes for each mouse. For a given gene the red and blue coloring indicates increased and decreased expression in P. chabaudi compared to naïve respectively. showed that both up and down regulated DE genes in the EM gd T cell gene set positively correlated with the DE genes in the CD8 + memory T cell gene set (P= 0.008; Figure 6C). To investigate if preexposed gd T cells could alter the course of infection in naïve mice, we isolated gd T cells from the spleen of either naïve or previously P. chabaudi exposed mice. These cells were then adoptively transferred into recipient C57BL/6 or RAG-1 mice (lacking T and B cells) that had been infected 1 day before with P. chabaudi iRBC. Additional control mice were infected and then injected with PBS. Parasitemia was measured daily by thin blood smears. We observed no significant difference in parasitemia or clearance of parasites between the experimental groups in the C57BL/6 WT mice (Supplementary Figure 2A). Assessment of whether pre-exposed gd T cells had a direct effect on infection in absence of adaptive immunity, yielded similar results with no significant change in parasitemia in RAG-1 mice between the groups and all mice were unable to control the infection (Supplementary Figure 2B). Altogether, these observations supports the novel concept that Plasmodium exposure induces EM gd T cells with a transcriptional profile resembling conventional memory T cells, but their protective role during a secondary infection in vivo remains to be determined.

DISCUSSION
In this study we used a malaria infection model to understand whether "memory-like imprints" were detectable in gd T cells after the infection was cleared and whether this was associated with memory-like gd T cell responses. We found that the transcriptional profile in pre-exposed EM gd T cells was significantly different from EM gd T cells from naïve mice and that differentially expressed genes in the pre-exposed EM gd T cells were positively correlated with previously reported differentially expressed genes in resting CD8 + memory T cells. Although the overlapping differentially expressed genes were not unique to T cells, elevated transcript levels of effector molecule genes in otherwise resting gd T cells were suggestive of an inherent functional ready-state that is a characteristic of conventional memory T cells. Furthermore, this showed that although gd T cell populations in both naïve mice and previously P. chabaudi-infected mice were classified as "memory" populations based on traditional surface markers, only pre-exposed gd T cells were observed to resemble that of conventional T cell memory. Consistent with their memory-like transcriptional profile, we also found that pre-exposure to antigen resulted in enhanced functional capacity of responding gd T cells upon encounter with cognate antigen. It has been suggested that as gd T cells emerge from the thymus, they have already acquired a functional imprint, which limits their plasticity in the periphery (46)(47)(48). Furthermore, functionally distinct gd T cells seem to have specific tissue distribution where spleen-derived gd T cells are predominately prone to producing IFNg (46). We found that pre-exposed gd T cells were multifunctional as they produced both IFNg and were cytotoxically active. However, the EM gd T cell population previously-exposed to malaria displayed significant reductions in the expression of genes associated with IL-17 responses, suggesting limitation to their functional plasticity after Plasmodium infection. Apart from low gene expression of IL-17a, this included significantly lower expression levels of Sox13 and Il1r1 genes. Sox13 is a lineage specific gd T cell transcription factor (49), which promotes IL-17 producing gd T cells (50) and IL-1 has recently been indicated to play an important role in supporting IL-17 production by antigen-specific T cells in vivo. Cells from Il1r1-deficient mice had dramatically reduced IL-17 production compared to cells from wild-type mice (51). Furthermore IL-17 producing gd T cells have been shown to rapidly respond to IL-23, which induces and supports IL-17 production (52)(53)(54). Interestingly following Plasmodium exposure, EM gd T cells have down regulated their Il-23r gene expression suggesting that they are less responsive to endogenous IL-23. As stimulation in vitro was carried out on bulk splenocyte preparations, which include CD4 + and CD8 + T cells, we could not exclude that a bystander effect was contributing to the activation of pre-exposed gd T cells. However, the transcriptional data from ex vivo EM gd T cells indicates that the functionally intrinsic characteristics of these cells is altered with infection and is maintained in absence of parasites.
We showed here that the gd T cell population in the spleen not only acquires memory-like characteristics, but also potentially fill an additional role as APCs. Although antigenpresentation and processing by gd T cells has previous been described, this characteristic remains relatively unexplored. This function is seemingly acquired upon TCR activation and human Vd2 T cells activated with the phosphoantigen isopentenyl A B C FIGURE 6 | Differentially expressed genes in P. chabaudi pre-exposed gd T cells are positively correlated with differentially expressed genes in CD8 + memory T cells.
(A) Venn diagram showing the number of overlapping and non-overlapping up-regulated (red) and down regulated (blue) genes. (B) Heatmap of the gene expression relative to P. chabaudi pre-exposed gd T cells data for the genes commonly significantly regulated (overlapping DE genes) between P. chabaudi pre-exposed gd T cells data and CD8 + memory T cell data. Each vertical column represents genes for each mouse. For a given gene the red and blue coloring indicates increased and decreased expression in P. chabaudi compared to naïve respectively. (C) Barcodeplot for the enrichment of DE genes in the resting CD8 + memory T cell data in P. chabaudi compared to naïve in the P. chabaudi pre-exposed gd T cell data, along with the ROAST p-value for the gene set testing.
pyrophosphate induced high levels of APC-related molecules, which resulted in a functional capacity to present antigens to ab T cells (55). In P. falciparum-infected individuals there is an increase of Vg9Vd2 T cells that express APC-related surface markers and this expression was induced by iRBCs (56). These cells were also able to elicit ab T cell responses in vitro suggesting that gd T cells may simply supplement existing APC populations. However, spleenderived gd T cells reside in an organ that plays a central role in the capacity to control and clear parasites and are in a location that allows them to encounter and remove blood-borne antigens and also initiate innate and adaptive immune responses. It is possible that following an initial malaria infection once an adaptive memory has been established, exposed gd T cells promote specific adaptive T cell functions. In support of this proposition, intestinal gd T cells have been found to have APC function and elicit distinct CD4 + T cell responses compared to responses induced by typical professional APCs (57). While the Vg9Vd2 T cell subset in humans are responsive to phosphoantigens (58), there is no evidence that murine gd T cells are equally responsive to this stimulation. Despite this, it is interesting to note that the APClike state of gd T cells show that similar induction occurred in vivo in mice. Furthermore, in contrast to Howard et al. (56), our study also demonstrated that this APC-relevant expression by gd T cells remained after clearance of the infection. However, a comprehensive understanding of the APC-capacities of tissueresident gd T cells and the specific functions that they provide for subsequent Plasmodium infections remains to be determined.
The work presented here demonstrates that blood-stage Plasmodium infection has a profound effect on the splenic gd T cell population, modifying its response capacity and gene expression profile. While our observations here support the existence of traditional memory cells with augmented secondary responses upon antigen re-encounter, their protective role during a secondary infection in vivo remains to be resolved. Evidence suggests that the role of gd T cells in protection is an intricate balance of timing, accessory signals from other immune cells, and also regulation of other immune cells by gd T cells (18,22,(59)(60)(61)(62). Collectively, this confounds effective investigation of protective capacity of memory-like gd T cells in a traditional sense using adoptive transfers into naïve mice upon challenge. Survival and effector capacity of preexposed gd T cells may be closely tied to other memory populations, which would not be present in newly infected mice. Therefore we were unable to determine whether preexposed memory-like gd T cell effector functions such as enhanced cytokine production and cytotoxic activity, could alter the course of infection or not upon antigen re-exposure in vivo. Furthermore, it also appears that effector function may not necessarily be the only role for these cells. Our findings here suggest a model by which antigen-experienced gd T cells undergo transcriptional changes that allows them to fulfil a novel role as antigen-presenting cells in subsequent infections. These findings have important implications for our understanding of the role of gd T cells in host immunity and gives insight into potential therapeutic modulations that can be achieved.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

AUTHOR CONTRIBUTIONS
RK and LI performed experiments and critically reviewed the manuscript. WA and DP analyzed data and critically reviewed the manuscript. DW, RM, and SS analyzed data. IM, DH, and LS provided conceptual input into the study design and critically reviewed the manuscript. EE conceived and performed experiments, analyzed data, and prepared the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by NHMRC grant APP106722 (EE). This work was made possible through Victorian State Government Operational Infrastructure Support and Australian Government NHMRC IRIISS. IM is supported by an NHMRC Senior Research Fellowship (#1043345). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.