NKG7 has a protective role during non-lethal experimental malaria following Plasmodium yoelii 17XNL infection

We previously showed a key role for NKG7 expressed by CD8⁺ T cells in the pathogenesis of ECM caused by infection of C57BL/6 mice with P. berghei ANKA (PbA) [53]. A key limitation of the ECM model is the acute and lethal nature of PbA infection. Therefore, to investigate the role of NKG7 in parasite control and the development of immunity against malaria, an experimental mouse model of non-lethal, resolving malaria caused by P. yoelii (Py17XNL) infection was employed. We first infected Nkg7-deficient (B6^ΔNkg) and wild-type (WT) control C57BL/6 mice, and measured the progression of blood parasitemia through the course of infection (Fig 1A). Higher frequencies of parasitised red blood cells (pRBCs) were observed in Nkg7-deficient mice, compared to their WT counterparts (Fig 1B). To assess the overall parasite burden more comprehensively, the area under the curve (AUC) was calculated from blood parasitemia data points, revealing a significant increase in parasite burden among Nkg7-deficient mice (P < 0.05; Fig 1C). These results indicate an important role for NKG7 in the control of Py17XNL infection.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. NKG7 has a protective role during non-lethal experimental malaria caused by Plasmodium yoelii 17XNL (Py17XNL) infection.

(A) C57BL/6 and B6.Nkg7^-/- (B6^∆Nkg7) mice were infected with non-lethal Py17XNL. Layout for the experimental timeline. (B) Frequency of parasitised RBCs (pRBC) in peripheral blood of B6 and B6^∆Nkg7mice (B). (C) Area Under the Curve (AUC), as a measure of parasite burden over the course of infection. Data is pooled from two independent experiments. A Mann-Whitney test was used to test for statistical significance, where ** P < 0.01. n = 11 per group.

https://doi.org/10.1371/journal.ppat.1013528.g001

The role of NKG7 in Th1 and Tfh cell responses during non-lethal experimental malaria following Plasmodium yoelii 17XNL infection

To investigate whether NKG7 played a role in CD4⁺ T cell development in this malaria model, Th1 and Tfh cell responses were examined at two specific time points during infection. Both Nkg7-deficient mice and WT mice were infected with Py17XNL, and flow cytometry analysis was conducted on days 7 and 14 post-infection (p.i.) (Fig 2A and 2B). We found no significant differences in body weight, development of splenomegaly or increase in spleen cell numbers between Nkg7-deficient mice and WT mice at either time point during infection (Fig 2C). Upon treatment with PMA and ionomycin in the presence of monensin, significantly decreased frequencies of Th1 (IFNγ⁺ T-BET⁺) and Tr1 (IFNγ⁺ IL-10⁺) cells were observed in Nkg7-deficient mice at day 7 p.i., when compared to WT mice (P < 0.05; Fig 2D). Additionally, significantly reduced frequencies and numbers of Tfh (PD-1⁺ CXCR5⁺) cells were noted at day 14 p.i. in Nkg7-deficient mice compared to their WT counterparts (P< 0.05; Fig 2D). A limitation with this data was no FoxP3 staining of CD4⁺ T cells, meaning that natural regulatory T (Treg) cells were not excluded from analysis. To determine whether they did contribute to the above results, we repeated these experiments and employed FoxP3 staining of CD4⁺ T cells (S1 Fig). These data showed no difference in the frequency or number of Treg cells between Nkg7-deficient mice and WT mice following Py17XNL infection (S1B-D Fig). Furthermore, no cytokine production was detected from these cells during infection (S1B and S1D Fig). Thus, these data suggest a potential role for NKG7 in Th1 and Tr1 cell development during the first week of Py17XNL infection, as well as in Tfh cell development during the second week of Py17XNL infection.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. The role of NKG7 in CD4⁺ T cells during non-lethal experimental malaria caused by Plasmodium yoelii 17XNL (Py17XNL) infection.

(A) Schematic showing the experimental timeline. Created in BioRender. Engwerda, C. (2026) https://BioRender.com/wxzptf9. (B) Gating strategy used to identify CD4⁺ T cell subsets. (C) Body and spleen weights of naïve and Py17XNL infected C57BL/6 and B6.Nkg7^-/- (B6^∆Nkg7) mice. The right graph shows the total splenocyte numbers of naïve and infected C57BL/6 and B6^∆Nkg7 mice. (D) Frequencies of IFNγ⁺ T-bet⁺ (Th1) cells and IFNγ⁺ IL-10⁺ (Tr1) cells as a percentage of TCRβ⁺ CD4⁺ T cells after stimulation with phorbol 12myristate 13-acetate and ionomycin. The corresponding numbers of these CD4⁺ T cell subsets in the spleen are also shown. The graphs on the right indicate frequencies of PD-1⁺ CXCR5⁺ (Tfh) cells as a frequency of TCRβ⁺ CD4⁺ T cells, and total number shown below. Data are from one representative experiment of two performed. A two-way ANOVA with Sidak’s multiple comparisons test was used to calculate statistical significance, where ** P < 0.01, *** P < 0.001, **** P < 0.0001. n = 6 mice per group.

https://doi.org/10.1371/journal.ppat.1013528.g002

These findings were supported when we examined the role of NKG7 in antigen-specific CD4⁺ T cells activation and differentiation in the lethal ECM model using PbT-II T cells. These are MHC-II-restricted TCR transgenic T cells which recognise a peptide derived from the PbA Hsp90 protein which is expressed by all rodent Plasmodium species [78]. To investigate cell-intrinsic roles for NKG7 in these cells, equal numbers of control (PbT-II^WT) and Nkg7-deficient (PbT-II^ΔNkg7) cells were adoptively transferred into congenic Ptprca (CD45.1) recipient mice prior to PbA infection. Cellular analysis was performed when the mice developed ECM at day 5 p.i. (S2A-B Fig). After treatment with PMA and ionomycin, in the presence of monensin, despite a small, but significant increase in IFNγ⁺ PbT-II^ΔNkg7 cells, there was a decrease in Tr1 (IFNγ⁺ IL-10⁺) cell frequencies, compared with PbT-II^WT cells from the same animals (P < 0.05; S2C Fig). PbT-II^ΔNkg7 cells also exhibited significantly lower frequencies of granzyme B-producing cells when compared to PbT-II^WT cells (P < 0.05), but no difference in CD107a expression (S2C Fig), as previously observed for Nkg7-deficient CD8⁺ T cells [53]. We next examined their functional potential by assessing expression of different co-inhibitory receptors, but no differences in the expression of these molecules by PbT-II^WT and PbT-II^ΔNkg7 Tr1 cells was found (S2D Fig). However, similar to granzyme B, a significant reduction in the frequency of perforin-producing PbT-II^ΔNkg7 cells was observed, when compared to the PbT-II^WT cells (P < 0.05; S2E Fig). Together, these results identify roles for NKG7 in CD4⁺ T cells during PbA infection in perforin and granzyme B expression, as well as the development of Tr1 cells.

Increased Nkg7 expression by Th1 and Tr1 cells during Plasmodium yoelii 17XNL infection

To characterise the cellular expression of NKG7 during Py17XNL infection, a Nkg7-transcriptional reporter mouse was utilised (Fig 3A-B). In these mice, cells in which Nkg7 transcription is initiated switch from expressing TdTomato (red) to GFP (green) [53]. In the naïve state, the majority of NK cells (Fig 3C) expressed GFP, but this expression significantly decreased following infection (P < 0.05), as previously reported in experimental visceral leishmaniasis [53]. Conversely, the frequency of GFP⁺ cells increased significantly in NKT, CD4⁺ and CD8⁺ T cell subsets during the infection (P < 0.05; Fig 3C). The frequency of GFP-expressing CD4⁺ T cells increased significantly from the naïve state to day 7 p.i. and was sustained at day 14 p.i. (P < 0.001 and P < 0.0001, respectively; Fig 3C). To further investigate this increase, we next examined Nkg7 expression within distinct CD4⁺ T cell subsets. Following activation with PMA and ionomycin, Th1 cells dominated in the naïve state, while Tr1 cells became dominant at days 7 and 14 p.i. (Fig 3D). The compartment representing other CD4⁺ T cells included cells that were neither Th1 cells, Tr1 cells, nor Treg cells. When analysing T cell subsets directly ex vivo, without further stimulation, a similar trend emerged with significantly increasing frequencies of GFP⁺ cells during the infection in CD4⁺ and CD8⁺ T cells (P < 0.05; Fig 3E), as well as in the CD4⁺ T cell subsets, including Th0 cells (T-BET^- CXCR6^-), Th1 cells (T-BET⁺ CXCR6⁺), and other CD4⁺ T cells (non-Tfh T-BET⁺ CXCR6⁺) (P < 0.05; Fig 3E). We were unable to detect IL-10-producing Tr1 cells without ex vivo activation with mitogens in this analysis. Interestingly, the frequency of GFP⁺ Tfh cells (PD-1⁺ CXCR5⁺) remained consistent between the naïve state and both infection time points (Fig 3E). These data suggest that NKG7 is expressed across multiple CD4⁺ T cell subsets with enrichment in Th1 and Tr1 cells following Py17XNL infection.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Nkg7 expression increases in Th1 and Tr1 cells during non-lethal Plasmodium yoelii 17XNL (Py17XNL) infection.

(A) Schematic showing the experimental timeline for B6.Nkg7-reporter mice (Nkg7-cre x mT/mG) infected with Py17XNL. Created in BioRender. Engwerda, C. (2026) https://BioRender.com/wxzptf9. (B) Gating strategy used to identify lymphoid cell subsets. (C) Frequencies of GFP⁺ (Nkg7⁺) in lymphoid cell subsets, as indicated, during Py17XNL infection. (D) Representative gating strategy and frequencies of GFP⁺ (Nkg7⁺) CD4⁺ IFNγ⁺ T-bet⁺ (Th1) cells, IFNγ⁺ IL-10⁺ (Tr1) cells, Foxp3⁺ regulatory T (Treg) cells and the remaining CD4⁺ T cells after stimulation with phorbal 12myristate 13-acetate and ionomycin, shown as a pie chart. The size of the compartments within the pie chart indicates the percentage of the CD4⁺ T cell subsets specified, within the GFP⁺ CD4⁺ T cell population. (E) Frequencies of all GFP⁺ (Nkg7⁺) cells, CD8⁺ T cells and CD4⁺ T cells during Py17XNL infection directly ex vivo. CD4⁺ T cells were further divided into CXCR6^- T-bet^- (Th0) cells, CXCR6⁺ T-bet⁺ (Th1) cells, PD-1⁺ CXCR5⁺ (Tfh) cells and the remaining CD4⁺ T cells, shown as a pie chart. The size of the compartments within the pie chart indicates the percentage of the CD4⁺ T cell subsets specified, within the GFP⁺ CD4⁺ T cell population. Statistical significance was assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test, where * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. n = 4 naïve or infected mice. Data are from one representative experiment of two performed.

https://doi.org/10.1371/journal.ppat.1013528.g003

A cell intrinsic role for NKG7 in the development of Th1, Tr1 and Tfh cells during Plasmodium yoelii 17XNL infection

To investigate whether NKG7 played a cell-intrinsic role in CD4⁺ T cells in the non-lethal malaria model, equal numbers of PbT-II^WT and PbT-II^ΔNkg7 cells were adoptively transferred into congenic Ptprca (CD45.1) recipient mice prior to Py17XNL infection. Cellular analyses were performed at both day 7 and day 14 p.i. to evaluate Th1, Tr1 and Tfh cell responses, with FoxP3 included in the staining panels as an exclusion marker (Fig 4A and 4B). In this adoptive transfer model where PbT-II^WT and PbT-II^ΔNkg7 cells were primed and expanded in the same host, we found greater expansion of PbT-II^ΔNkg7 cells, relative to PbT-II^WT cells during the first seven days of infection (Fig 4B), suggesting a CD4⁺ T cell cell intrinsic role for NKG7 in cell expansion not observed in Nkg7-deficient mice (Fig 2D). However, at day 7 p.i., following treatment with PMA and ionomycin in the presence of monensin to assess cellular potential, no discernible differences were observed in the frequencies of Th1 cells (IFNγ⁺ T-BET⁺) between the two PbT-II populations (Fig 4C). We did observe a significant reduction in the frequencies of PbT-II^ΔNkg7 Tr1 cells (IFNγ⁺ IL-10⁺) (P < 0.05; Fig 4C). In unstimulated cells analysed ex vivo, a significant decrease was noted in PbT-II^ΔNkg7 Tfh cell (PD-1⁺ CXCR5⁺) frequencies, as well as PbT-II^ΔNkg7 Th1 (T-BET⁺ CXCR6⁺) cells (P < 0.05; Fig 4C). CXCR6 was used because it was previously identified as a robust Th1 cell marker of PbT-II cells without the need to ex vivo stimulation [79,80]. At day 14 p.i., in unstimulated cells, significant differences were only observed in Th1 (T-BET⁺ CXCR6⁺) cells (P < 0.05; Fig 4D). Taken together, these data provide support for a cell-intrinsic role of NKG7 in the development of Th1, Tr1 and Tfh cells at day 7 p.i. and Th1 cells at day 14 p.i..

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. NKG7 has a cell intrinsic role in CD4⁺ T cell responses during non-lethal experimental malaria caused by Plasmodium yoelii 17XNL (Py17XNL) infection.

(A) Experimental timeline for co-transfer of transgenic PbT-II^WT and PbT-II^ΔNkg7 cells into B6.CD45.1 mice prior to infection with Py17XNL. Created in BioRender. Engwerda, C. (2026) https://BioRender.com/wxzptf9. (B) Gating strategy used to identify and phenotype PbT-II cell subsets with or without stimulation with phorbol ester and ionophore. Frequency of PbT-II cells, relative to conventional (cCD4), as well as total PbT-II cell numbers on day 7 post-infection. (C) Frequencies of co-transferred transgenic PbT-II^WT and PbT-II^ΔNkg7 IFNγ⁺ T-bet⁺ (Th1) cells and IFNγ⁺ IL-10⁺ (Tr1) cells as a proportion of parental PbT-II cells (CD45.1/2+ (PbT-II^WT) and CD45.2+ (PbT-II^ΔNkg7)) from the spleen at day 7 post infection after stimulation with phorbol ester and ionophore (two left panels). Frequencies of co-transferred transgenic PbT-II^WT and PbT-II^ΔNkg7 T-bet⁺ CXCR6⁺ cells and PD-1⁺ CXCR5⁺ (Tfh) cells as a proportion of parental PbT-II cells, as above, from the spleen at day 7 post infection ex vivo (no further stimulation; two right panels). (D) Frequency of the same PbT-II cells as above at day 14 post infection. A Wilcoxon test was used to define statistical significance, where * P < 0.05, n = 6 mice per group. Data are from one representative experiment of two performed.

https://doi.org/10.1371/journal.ppat.1013528.g004

To test whether the above results could be caused by NKG7-dependent changes in proliferation or apoptosis, an early time course assessment of PbT-II^WT and PbT-II^ΔNkg7 cell proliferation, activation and apoptosis was performed. The PbA model rather than the Py17XNL model was chosen to address this question because it provides more rapid and robust PbT-II cell proliferation over a short timeframe. Equal numbers of PbT-II^WT and PbT-II^ΔNkg7 cells were adoptively transferred into congenic Ptprca (CD45.1) recipient mice prior to PbA infection and cellular analysis performed each day, starting from the day of infection to day 5 p.i. (S3A-B Fig). We used this more acute infection model because it provides rapid and robust PbT-II activation of a short time frame. Expansion of both PbT-II cell populations was observed at day 4 p.i. (S3C Fig), although the frequency of PbT-II^ΔNkg7 cells, was reduced at days 4 and 5 p.i., relative to PbT-II^WT cells. Analysis of CTV dilution revealed limited differences between PbT-II^WT and PbT-II^ΔNkg7 cells, and both populations had undergone approximately six rounds of cell division by day 3 p.i. (S3D Fig), with most cells CTV-negative by day 4 p.i.. In addition, few differences in CD4⁺ T cell-specific activation markers between PbT-II^WT and PbT-II^ΔNkg7 cells were seen, although an early increase in the frequencies of CD69⁺ and CD25⁺ PbT-II^ΔNkg7 cells, as well as small changes in CD11a expression, relative to PbT-II^WT cells was noted one day after infection for CD69 and CD25, and throughout infection for CD11a (S3E Fig). Apoptosis analysis of PbT-II^WT and PbT-II^ΔNkg7 cells demonstrated limited differences in frequencies of early apoptotic (Annexin V⁺ 7-AAD^-) PbT-II^ΔNkg7 cells, compared to PbT-II^WT cells, and no differences in frequencies of cells that were in late apoptosis or already dead (Annexin V⁺ 7-AAD⁺) (S3F Fig). Together, these results show limited roles for NKG7 in proliferation and apoptosis following PbA infection, although the reduced expansion of PbT-II^ΔNkg7 cells, relative to PbT-II^WT cells at days 4 and 5 p.i., accompanied by limited changes in proliferation and apoptosis, indicate potential model and timepoint limitations compared with Py17XNL infection, possibly reflecting tissue sequestration of CD4⁺ T cells in the former model.

The expression of NKG7 during controlled human infection studies with Plasmodium falciparum

We next examined the pattern of NKG7 expression by human immune cells during malaria. To measure NKG7 expression in human peripheral blood mononuclear cells (PBMCs), we first tested the specificity of a previously reported monoclonal antibody [58]. We employed HEK293 cells transfected with human NKG7-GFP and the NK cell line NK92, for which CRISPR/cas9 was employed to delete NKG7. We readily detected NKG7 in NKG7-transfected HEK293 cells, but not in control cells, as well as NK92 wild-type cells, but not cells depleted of NKG7, where expression was significantly diminished at both mRNA and protein levels (S4 Fig). Thus, this anti-NKG7 antibody was included in flow cytometry antibody panels for further investigations.

We next examined blood from healthy, malaria-naïve volunteers participating in controlled human malaria infection (CHMI) studies with blood-stage P. falciparum[81]. This involved samples from 8 individuals (5/8 male, aged median 22, range 18–29); all were treated with an anti-parasitic drug 8–9 days after infection, when blood parasite level exceeded 5000 parasites/mL in most participants [82]. PBMCs were isolated from blood and examined before infection (day 0), at pre-treatment (day 8 p.i., prior to drug administration), after anti-parasitic treatment (day 15 p.i.) and at the end of the study (EOS; day 45 ± 2) (Fig 5A). Unsupervised dimensionality reduction using UMAP, followed by clustering of flow cytometry data, identified 30 distinct cell clusters, (Figs 5B and S5A). Cell clusters with high NKG7 expression also had high expression of the cytotoxic molecules GZMB, CD107a and KLRG1 (Fig 5C). We identified one CD4⁺ T cell cluster (C04) and five CD8⁺ T cell clusters (C07, C13, C14, C15 and C17) that had high expression of NKG7 (Fig 5D-G). The NKG7⁺ CD4⁺ T cell cluster (C04) displayed characteristics of an effector memory CD4⁺ T cell cluster (CD25^- CD127⁺ CCR7^-) that expressed multiple activation markers, including CD49b, PD1, CD38, KLRG1, GZMB and CD107a (Fig 5D and F). The five NKG7⁺ CD8⁺ T cell clusters lacked CCR7 expression and were identified as likely effector memory T cell clusters with varying levels of CD45RA expression. Three of the NKG7⁺ CD8⁺ T cell clusters (C07, C13 and C14) showed a similar cytotoxic phenotype with different combinations of KLRG1, CD38, and PD1 expression. Cluster 17 appeared to be a less differentiated CD8⁺ T cell cluster as it lacked expression of cytotoxic, activation or exhaustion markers and had low expression of KLRG1 and CCR7. Cluster 15 showed differentiation from the other four NKG7⁺ CD8⁺ T cell clusters and expressed high levels of CCR6, CD56 and KLRG1 and lacked expression of cytotoxic molecules (Fig 5D and G), suggesting it may represent either innate-like CD8⁺ T cells [83] or non-cytotoxic memory-like CD8⁺ T cells [84]. Examination of the differential abundance of cell clusters after infection (day 0 to day 8, day 0 to day 15, day 0 to EOS), revealed a significant increase (P < 0.05) in the abundance of the one NKG7⁺ CD4⁺ T cell cluster (C04) and one of the NKG7⁺ CD8⁺ T cell clusters (C07) when day 0 (pre-infection) was compared to day 15 (15 days p.i., 7 days post-drug treatment) (Fig 5E).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. NKG7 expression correlated with cytotoxic and activated cells during controlled human malaria infection (CHMI).

(A) Schematic showing the experimental design for CHMI studies; malaria-naïve subjects were infected with blood-stage Plasmodium falciparum 3D7 followed by treatment with an anti-parasitic trial drug, 8-9 days after infection. Peripheral blood mononuclear cells (PBMCs) were isolated from blood samples that were collected for analysis before infection (day 0), pre-treatment (day 8 post infection (p.i.), prior to drug administration), post anti-parasitic treatment (day 15 p.i.) and at the end of the study (EOS). Created in BioRender. Engwerda, C. (2026) https://BioRender.com/wxzptf9. (B) Cryopreserved PBMCs were thawed and immunophenotyped by spectral flow cytometry. Post-acquisition analysis was conducted using OMIQ software for dimensionality reduction using unsupervised FLOWSOM clustering that yielded 30 clusters. (C) Mean expression of cytotoxic markers NKG7, granzyme B (GzmB), CD107a, and KLRG1 are also shown separately on the UMAP. (D) Expression of cell phenotyping markers were measured, and mean fluorescence intensity (MFI) of staining presented as a heatmap. n = 31 volunteer samples, day 0 (n = 8), day 8 (n = 8), day 15 (n = 8), EOS (n = 7). (E) PBMCs were divided into 30 clusters and clusters that were significantly different in abundance between day 0 and 8, day 0 and 15, and day 0 and end of study (EOS) are indicated with red circles for significant increase and blue for significant decrease (significance assessed using EdgeR). T cell clusters of interest are underlined. (F) Frequency (top) and Dot plots (bottom) depicting Cluster change, expression patterns and frequency of positive events for cytotoxic and activation markers in a CD4⁺ T cell cluster and (G) CD8⁺ T cell clusters for the duration of the CHMI trial.

https://doi.org/10.1371/journal.ppat.1013528.g005

Further analysis of samples at day 15 p.i., when T cell activation appeared to peak, revealed significantly increased expression of CD38, PD1 and HLA-DR within the NKG7⁺ CD4⁺ T cell cluster (C04) (P < 0.05; Figs 5F and S4B). However, significantly decreased NKG7 expression from day 0 to day 15 was observed within this population, mirroring a decrease in CD107a expression (P < 0.05), and supporting a role for NKG7 in degranulation of effector molecules (S5B Fig). Cluster 7, which was the only CD8⁺ T cell cluster that was significantly different in abundance from day 0 to day 15 p.i. (Fig 5E), had relatively high expression of CD38, HLA-DR and PD-1 (Fig 5G), suggesting a highly activated and exhausted cluster. Similar to the NKG7⁺ CD4⁺ T cell cluster C04, a significantly decreased expression of CD107a from day 0 to day 15 was also observed in all NKG7⁺ CD8⁺ T cell clusters (P < 0.05; Figs 5G and S5B Fig). Together, these data provide new insights into CD4⁺ and CD8⁺ NKG7-expressing T cells during CHMI with multiple T cell subsets identified, each with a different functional potential.

Species-specific differences in NKG7-expressing T cells during uncomplicated Plasmodium falciparum and P. vivax malaria

To test whether the NKG7 expression patterns observed in CHMI samples were also found during naturally acquired infection, we next investigated blood samples from patients admitted to a tertiary-referral hospital in Sabah, Malaysia, during 2010–2012, with uncomplicated P. falciparum or P. vivax malaria. Blood was collected at the time of presentation and one month post anti-parasitic treatment. These samples were classified into three groups: acute P. falciparum (Pf) infection (n = 5), acute P. vivax (Pv) infection (n = 6) and convalescence. The convalescence group included samples from 2 Pf and 6 Pv patients. Following analysis of flow cytometry data, 20 clusters were identified (Fig 6A), with the majority of NKG7-expressing clusters also expressing the cytotoxic molecules granulysin (GNLY) and/or granzyme B (GZMB; Fig 6B). Of the NKG7⁺ clusters, two CD4⁺ T cell clusters and four CD8⁺ T cell clusters were identified (Fig 6B and C). The two NKG7⁺ CD4⁺ T cell clusters (C05 and C20) appeared to be effector memory cell clusters as both had minimal CD27 and CD45RA expression (Fig 6B), and cluster 20 having high CD57 expression (Fig 6C). Three of the NKG7⁺ CD8⁺ T cell clusters were grouped together in the UMAP (C11, C17 and C18), with cluster 11 and cluster 17 overlapping, suggesting phenotypic and possible functional similarities (Fig 6A). Cluster 13 was clearly separated from the other three NKG7⁺ CD8⁺ T cells clusters in the UMAP (Fig 6A) and had high CD161 expression (Fig 6B and C), overlapping several NKG7⁺ CD8⁺ T cells clusters observed in CHMI samples (Fig 5).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Differences in NKG7 expression during P. falciparum and P. vivax naturally acquired malaria.

(A) Peripheral blood mononuclear cells (PBMCs) were isolated from blood samples collected from patients with naturally acquired, uncomplicated malaria at the Queen Elizabeth Hospital in Sabah, acute P. falciparum (Pf; n = 5) and acute P. vivax (Pv; n = 6). Post-treatment samples were also collected (n = 8). Cryopreserved PBMCs were thawed and immunophenotyped by spectral flow cytometry. Post-acquisition analysis was conducted using OMIQ software for dimensionality reduction using and unsupervised FLOWSOM clustering. (B) Expression of cell protein markers was measured and mean fluorescent intensity of staining presented as a heatmap. (C) Samples were divided into 20 clusters and the frequency of CD4⁺ and CD8⁺ T cell clusters are shown (top) and dot plots (bottom) depicting Cluster change, expression patterns and frequency of positive events for cytotoxic and activation markers compared between Pf and Pv. (D) NKG7 expression for CD4⁺ T cell clusters, and (E) HLA-DR, CD162 and KLRG1 expression for all CD8⁺ T cell clusters are shown. Significance was assessed by ordinary one-way ANOVA with Tukey’s multiple comparisons test, where * P < 0.05, n = 5-6 patient samples per group.

https://doi.org/10.1371/journal.ppat.1013528.g006

By using the same analysis methods employed for CHMI samples, the two NKG7⁺ CD4⁺ T cell clusters (C05 and C20) were found to be the only clusters with significant differences in abundance when comparing the acute infections of the two different Plasmodium species, with both clusters increased during acute Pf infection relative to Pv (S4 Fig). When acute Pf infection was compared to convalescence, none of the four significantly different clusters included CD4⁺ or CD8⁺ T cells (S5 Fig), although two NKG7-expressing NK cell clusters were increased (clusters 7 and 8; S6 Fig). However, one of the four clusters identified as being significantly increased in frequency in acute Pv infection, compared to convalescence, was cluster 11, a NKG7⁺ CD8⁺ T cell cluster, consistent with a role for cytotoxic CD8⁺ T cells in immune responses to Pv malaria [47] (S6 Fig).

We also compared differences within the NKG7⁺ CD4⁺ and CD8⁺ T cells clusters between patient groups. In both NKG7⁺ CD4⁺ T cell clusters, there was significantly higher expression of NKG7 expression in acute Pv infection, when compared to acute Pf infection (P < 0.05; Fig 6D). The only other difference between the NKG7⁺ CD4⁺ T cell clusters was CD57 expression, which was much higher overall in cluster 20, although no difference was observed between patient groups. This cluster also had a small increase in the overall expression of cytotoxic molecules, particularly granzyme B and granulysin, relative to the other CD4⁺ T cell clusters, cluster 5 (Fig 6B-C). The four NKG7⁺ CD8⁺ T cell clusters also had increased NKG7 expression during acute Pv infection when compared to acute Pf infection but was only significantly increased in cluster 18 (P < 0.01; Fig 6D). Cluster 18 also had significantly higher expression of HLA-DR during Pv infection when compared to Pf infection (P < 0.05; Fig 6E) and overall had higher CD57 expression in both groups when compared to the other clusters. Cluster 13 was a non-cytotoxic cluster and had high levels of CD161 expression, which was significantly increased in acute Pf infection, when compared to acute Pv infection (P < 0.05; Fig 6E). KLRG1 expression in cluster 13 followed the same pattern when the two acute infections were compared (P < 0.05; Fig 6E). Overall, this data confirms that the NKG7 expression patterns found in CD4⁺ and CD8⁺ T cells in CHMI samples were similar to those observed in natural infection, with some notable differences between patients infected with different Plasmodium species.

Mice and human ethics

Mice. Experimental mice use was in accordance with the “Australian Code of Practice for the Care and Use of Animals for Scientific Purposes” (Australian National Health and Medical Research Council) and approved by the Queensland Institute of Medical Research (QIMR) Berghofer Medical Research Institute Animal Ethics Committee (AEC; AEC reference number P2304; approval number: A1707-615M).

CHMI subjects. Human ethics approval was provided by the QIMR Berghofer Medical Research Institute Human Research Ethics Committee (HREC; HREC reference number P1479). Written informed consent was received from all participants [82].

Endemic country malaria patients. Ethics approval for the use of human samples was obtained from the QIMR Berghofer Human Research Ethics Committee (HREC reference number P3444), the Northern Territory Department of Health and Menzies School of Health Research ethics committee (HREC 2010–1431), and Medical Research and Ethics Committee, Ministry of Health, Malaysia (NMRR-10-754-6684 and NMRR-12-499-1203). Written informed consent was obtained from all adult study participants or, in the case of children, parents or legal guardians. The inclusion criteria included being 12 years of age or older.

Experimental mice and human participant details

Mice. The experiments involved the use of female mice aged six week or older, unless otherwise specified. Mice were kept in group housing, with a maximum of six mice per cage. They were maintained in a pathogen-free environment at the QIMR Berghofer Medical Research Institute Animal Facility (Brisbane, Queensland, Australia). B6.Cd45.1 (Ptprca) mice were obtained from the Animal Resources Centre (Perth, Western Australia, Australia). C57BL/6J (WT) mice were obtained from the Walter and Eliza Hall Institute (Melbourne, Victoria, Australia). All other mice were bred at QIMR, including B6J.Nkg7^–/–, PbT-II, PbT-I and Nkg7-reporter (Nkg7-cre (C57BL/6J-Nkg7^{em1(cre)Wehi/}J; JAX stock #040147) × mT/mG) lines. Transgenic PbT-II mice [78] were bred with B6J.Nkg7^–/– mice to produce Nkg7-deficient PbT-II mice (PbT-II^ΔNkg7; CD45.2⁺ CD45.1^–) mice, and bred with B6.Cd45.1 mice to produce PbT-II × B6.Cd45.1 (PbT-II^WT; CD45.1⁺ CD45.2⁺). Transgenic PbT-I mice [108] were bred with B6J.Nkg7^–/– mice to produce Nkg7-deficient PbT-I mice (PbT-I^ΔNkg7; CD45.1^– CD45.2⁺) mice, and bred with B6.Cd45.1 mice to produce PbT-I × B6.Cd45.1 (PbT-I^WT; CD45.2⁺ CD45.1⁺).

CHMI subjects. PBMCs were obtained from volunteers who participated in a CHMI study involving P. falciparum. The primary investigation was a Phase 1b trial conducted at a single centre [82]. The trial aimed to evaluate the safety, tolerability, pharmacokinetic profile, and antimalarial activity of single doses of co-administered artefenomel and piperaquine phosphate against early P. falciparum blood-stage infection in healthy adult volunteers. Here, PBMCs from participants who consented for involvement in a study to analyse immunological responses were used. This included 8 individuals, 5/8 male, aged median 22, range 18–29.

Endemic country malaria patients. Blood was obtained from patients enrolled in a comparative clinical-pathophysiological study conducted at Queen Elizabeth Hospital (QEH) in Kota Kinabula, Sabah, Malaysia, spanning from September 2010 to October 2011 [109,110]. This study incorporates analysis of PBMCs obtained from patients recruited in the primary study, including 5 individuals diagnosed with uncomplicated P. falciparum (2/5 male, median age 39, range 36–54) and 6 with uncomplicated P. vivax malaria (4/6 male, median age 42, range 16–54).

Plasmodium infection in mice. Plasmodium infections in mice were established using parasites that had undergone a single passage in C57BL/6J mice. 200 μl of transgenic PbA (231c11) parasites, which were an in-house laboratory stock stored at −80 °C and expressing luciferase and GFP, were thawed at room temperature (RT). These parasites were then injected into a passage mouse via an intraperitoneal (i.p.) route. Py17XNL stabilates (in-house laboratory stock, frozen at −80 °C) were washed in NaCl and resuspended in 1 ml sterile PBS (1x) and 200 µl was injected via intravenous (i.v.) route into a passage mouse. Both the PbA and Py17XNL passage mice were monitored with parasitemia checks from day 3 post infection (p.i.) and sacrificed when >1% pRBCs. The passage mouse was euthanised and blood was obtained through cardiac puncture and diluted into RPMI/PS (containing 1 IU/ml heparin). The collected blood was then spun at 290 g for 7 min at RT. A hemocytometer (Pacific Laboratory Products) was used to enumerate RBCs. Parasite preparations were made, each containing 5 × 10⁵ pRBCs/ml. Subsequently, experimental mice were injected via i.v. route with 200 μl of this preparation, providing approximately 1 × 10⁵ pRBCs per mouse.

Quantification of Plasmodium parasitemia in mice. Passage and experimental mice were tail bled to collect two drops of blood into 250 μl RPMI/PS containing 1 IU/ml heparin. Next, 25 µl of this solution was incubated with SYTO 84 (Invitrogen; Life Technologies) and Hoechst 33342 (Sigma–Aldrich) in RPMI/PS for 22 min at RT. Then, each sample was quenched with 300 µl of RPMI/PS and acquired on a BD LSRFortessa through BD FlowJo version 10.0 (BD Biosciences).

Preparation of spleen single-cell suspensions. Mice were euthanized and a mid-sagittal incision was performed on the abdominal cavity and the spleen was carefully removed. The spleen’s weight was recorded, and it was then placed into 1% fetal bovine serum (FBS) in PBS (1% FBS/PBS). To isolate the cells, the spleens were mechanically mashed using the rear end of a 5 ml syringe plunger (Terumo Medical) through a 100 μm cell strainer (EASYstrainer; Greiner Bio-One). The resulting cell suspension was resuspended in 1% FBS/PBS. Using an Eppendorf Centrifuge 5810R (Thermo Fisher Scientific), the cell suspension was then subsequently centrifuged at 350 g. To remove RBCs, the cell pellet was lysed by treating and agitating for 7 min at RT in Red Blood Cell Lysing Buffer Hybri-Max (Sigma–Aldrich). For enumeration, the cells were diluted in Trypan Blue Stain (Invitrogen) and 1x DPBS (Gibco). Further method details included in [53].

Conventional and spectral flow cytometry. Flow cytometry staining protocols were carried out using Falcon 96-Well Clear Round Bottom Tissue Culture-Treated Cell Culture Microplates (Corning). For murine pre-stains, samples were incubated in 30 µl of complete media containing fluorescence-conjugated antibodies reactive against surface proteins for 30 min at 37 °C. 30 µl Zombie Aqua Fixable Viability Dye cocktail (BioLegend) or LIVE/DEAD Fixable Blue Dead Cell Stain Kit (Thermo Fisher) and TruStain FcX (BioLegend; anti-mouse CD16/32) were used for 15 min at RT to treat single-cell suspensions. During a 2 h stimulation at 37 °C, staining for CD107a (LAMP-1) was carried out with PMA/ionomycin or treatment with monensin (described later) by including 1 µl anti-Mouse CD107a (clone 1D4B from BioLegend) to 100 µl of the stimulation cocktail. Other PMA stimulations were carried out for a duration of 2.5 to 3 h. After treatment, cells were washed once with 1% FBS/PBS and spun at 575 g for 1 min at 4 °C in an Eppendorf Centrifuge 5810R (Thermo Fisher Scientific). Subsequently, 30 µl of a master mix containing antibodies, conjugated with various fluorophores, reactive against surface proteins were used to stain samples for 30 min at 37 °C. After two washes with 1% FBS/PBS staining buffer, samples were treated with 50 µl of fixation buffer for 20 min at RT. Fixation buffers used were from either the BD Cytofix Fixation Buffer Set (BD Biosciences) or the eBioscience Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific). Cells were then rinsed two times with wash buffers from the corresponding kits by spinning for 1 min at 953 g at RT. After this, cells were treated with 30 µl of a cocktail of antibodies, conjugated with various fluorophores, against intracellular proteins for 45 min at 37 °C. For human staining, 10⁶ PBMCs per participant, per time point, were aliquoted into plates. Samples were cultured with 50 µl of LIVE/DEAD Fixable Blue Dead Cell Stain Kit (diluted in 1x PBS; Thermo Fisher) at RT for 15 min. After this, cells were rinsed with 2% FBS/PBS and spun for 5 min in an Eppendorf Centrifuge 5810R (Thermo Fisher Scientific). Subsequently, 25 µl of a first master mix of antibodies, conjugated with various fluorophores, reactive against surface proteins samples were used to stain samples for 15 min at RT. After this, 50 µl of a second master mix of antibodies, conjugated with various fluorophores, reactive against surface proteins was added to PBMCs for a further 30 min at RT. After two washes with 2% FBS/PBS staining buffer, samples were treated with 100 µl of fixation buffer from the eBioscience Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific) for 20 min on ice. Cells were then washed twice with permeabilisation wash buffer. Following this, 50 µl of a master mix of antibodies, conjugated with various fluorophores, reactive against intracellular proteins cells were used to stain samples for 30 min on ice. All staining procedures were conducted in the dark. Prior to acquisition, samples were stored at 4 °C. Samples were acquired on a BD LSRFortessa (BD Biosciences) using BD FACSDiva version 8.0 or Cytek Aurora 5 laser using the SpectroFlo software package version 2.2 (Cytek Biosciences), then analysed on FlowJo version 10 OSX (FlowJo). GraphPad Prism 9 (version 9.4.0; GraphPad Software) was used for graphing and statistical analyses were conducted with statistical significance defined as P ≤ 0.05. Further method details included in [53].

For assessment of apoptosis and cell viability, cells were washed twice with PBS following surface staining, then diluted in 1x Annexing V binding buffer and 5ul of Annexing V-PE added together with 5ul of 7-AAD reagent were added. Cells were incubated for 15 minutes at room temperature, followed by washing once with 1x Annexing V binding buffer and analyzed by flow cytometry within 1 hour of staining.

Mouse flow cytometry markers used to define cell subsets can be found in Tables 1–5.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Phenotypic flow cytometry markers for mouse datasets (S1 Fig).

https://doi.org/10.1371/journal.ppat.1013528.t001

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Phenotypic flow cytometry markers for mouse datasets (Fig 2).

https://doi.org/10.1371/journal.ppat.1013528.t002

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Phenotypic flow cytometry markers for mouse datasets (Fig 3).

https://doi.org/10.1371/journal.ppat.1013528.t003

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Phenotypic flow cytometry markers for mouse datasets (Fig 4).

https://doi.org/10.1371/journal.ppat.1013528.t004

Download:

• PNG
  larger image
• TIFF
  original image

Table 5. Phenotypic flow cytometry markers for apoptosis, activation and proliferation markers (S2 Fig).

https://doi.org/10.1371/journal.ppat.1013528.t005

Human flow cytometry markers used to define cell subsets can be found in Tables 6 and 7.

Download:

• PNG
  larger image
• TIFF
  original image

Table 6. Phenotypic flow cytometry markers for human datasets (Figs 5 and 6).

https://doi.org/10.1371/journal.ppat.1013528.t006

Download:

• PNG
  larger image
• TIFF
  original image

Table 7. Cytotoxic flow cytometry markers for endemic dataset (Fig 6 and S4 Fig).

https://doi.org/10.1371/journal.ppat.1013528.t007

PMA/ionomycin restimulation. Cells were cultured in complete media supplemented with 25 ng/ml PMA (Sigma–Aldrich) and 1 μg/ml (1.33 nM) ionomycin calcium salt (Sigma–Aldrich), with the addition of 2x monensin solution (BioLegend). The PMA/ionomycin restimulation process was carried out over a 2–3 h duration at 37 °C in the presence of 5% CO₂. Further method details included in [53].

Isolation of PbT-II and PbT-I cells by FACS. PbT-II cells were enriched prior to FACS by using either the CD4 + T cell Isolation Kit, mouse (Miltenyi Biotec) or EasySep Mouse CD4 + T Cell Isolation Kit (STEMCELL) following the manufacturer’s guidelines. PbT-I cells were enriched prior to FACS by using either the CD8 + T cell Isolation Kit, mouse (Miltenyi Biotec) or EasySep Mouse CD8 + T Cell Isolation Kit (STEMCELL) following the manufacturer’s guidelines. PbT-II and PbT-I cells were stained for surface proteins prior to isolation by cell sorting for RNA-sequencing and real-time quantitative PCR (RT-qPCR). The surface staining included use of anti-mouse CD90.2 (BV605; 53-2.1), anti-mouse CD45.2 (PE; 104), anti-mouse CD45.1 (FITC; A20) and either anti-mouse CD4 (APC; GK1.5) for PbT-II cells or anti-mouse CD8 (APC; 53-6.7) for PbT-I cells (all from BioLegend). Dead cells were omitted during flow cytometry with a positive stain for SYTOX Blue Dead Cell Stain. Cells were sorted on a BD FACSAria II or BD FACSAria III (Becton Dickinson), treated in buffer RLT (Qiagen) and stored at −80 °C. Furhter method details included in [53].

Adoptive cell transfer of PbT-II and PbT-I cells. PbT-II cells were isolated from the spleens of PbT-II^WT (CD45.1 x PbT-II) and PbT-II^ΔNkg7 mice using either the CD4⁺ T cell Isolation Kit, mouse (Miltenyi Biotec) or EasySep Mouse CD4 + T Cell Isolation Kit (STEMCELL) following the manufacturer’s guidelines. PbT-I cells were isolated from the spleens of PbT-I^WT (CD45.1 x PbT-I) and PbT-I^ΔNkg7 mice using either the CD8a⁺ T Cell Isolation Kit, mouse (Miltenyi Biotec) or EasySep Mouse CD8⁺ T Cell Isolation Kit (STEMCELL) following the manufacturer’s guidelines. PbT-II^WT and PbT-II^ΔNkg7 cells were enumerated and pooled in equal numbers to a final concentration of 1x10⁶ cells/ml. Then, 1 day before infection with PbA or Py17XNL, 200 μl (2x10⁵ cells) was injected via i.v. route into the tail vein of B6.Cd45.1 recipient mice. Similarly, PbT-I^WT and PbT-I^ΔNkg7 cells were enumerated and pooled in equivalent numbers to a final concentration of 5x10⁶ cells/ml. Then, 200 μl (1x10⁶ cells) were injected via i.v. route into the tail vein of B6.Cd45.1 recipient mice 1 day before infection with PbA. Additional method details included in [53]. For assessment of PbT-II^WT and PbT-II^ΔNkg7 proliferation, cells were loaded with 5uM CTV (Violet Celltrace) for 15 minutes at room temperature and washed once before cell transfer.

HEK293T transient transfection. HEK293T cells were plated at 0.5 x10⁶ in 60mm plate (Falcon) with 2 ml DMEM (Gibco) supplemented with 10% heat inactivated FCS (Gibco) for 48 hours at 37°C 5% C0₂ or until 50–60% confluent. NKG7-GFP reporter cells were generated by adding a transfection mix of RD114 envelope plasmid (1.5 μg), PeqPam Gag-pol plasmid (2 μg) and vector plasmid eGFP-T2A-NKG7 (6.25 μg) with 19 μl FuGENE 6 transfection reagent (Promega) in a final volume of 470 μl of DMEM, dropwise to the cells with gentle mixing. After transfection cells were incubated for 72 hours before flow cytometry analysis.

NK92 Cell culture. NK92 human NK cell line (Malignant non-Hodgkin’s lymphoma) was cultured in RPM1640 (Gibco) supplemented with 10% heat inactivated FCS (Gibco), 1x Glutamax (Gibco), 100 units/ml Penicillin-Streptomycin (Gibco) and 1000 units/ml recombinant human IL-2 (Miltenyi). Cells were incubated at 37°C 5% C0₂ and were passaged every 48 hours in 10 ml volume at a density of 2–5 x10⁵ cells/ml.

Nucleofection and CRISPR Cas9. Gene editing was performed on 5 x10⁵ NK92 cells as previously described (Huang RS). Briefly, for CAS9-RNP preparation 2ul of CAS9- NLS (40 μM) (QB3 Macrolab) were mixed with 2.4ul of human NKG7 sgRNA sequence: CATGGCCACCACCATGGAGA (100 μM) (Synthego) and incubated for 15 minutes at 37°C. NK92 cells were washed once in dPBS (Gibco) and suspended in 20ul of freshly prepared nucleofection buffer containing 150 mM sodium phosphate buffer (pH 7.2), 5 mM KCl, 15 mM MgCl₂, 15 mM HEPES and 50 mM mannitol (SIGMA), cells were added to the CAS9-RNP construct and 24.4 μl volume was transferred to a 16 well nucleofection strip (Lonza) for pulsing with program CM-137 on the Amaxa 96 well shuttle (Lonza). Nucleofected cells were transferred to a 24 well plate (Falcon). One ml of warm culture media was added and cells were incubated for 96 hours at 37°C 5% C0₂, followed by flow cytometry staining.

RNA extraction. NK92 cells were preserved in RLT buffer and subsequently homogenised in QIAshredder columns (both from Qiagen). The RNA extraction process was performed using either the RNeasy Mini Kit or RNeasy Micro Kit (both from Qiagen), following the manufacturer’s guidelines. RNA integrity was assessed using the RNA 6000 Pico Kit (Agilent Technologies) (Performed at QIMR Berghofer sequencing facility). Subsequently, the isolated RNA for RT-qPCR was reverse transcribed into cDNA using the iScript cDNA Synthesis Kit (Bio-Rad) following the manufacturer’s instructions.

RT-qPCR. Taqman Gene Expression Assays (Life Technologies) were utilised in conjunction with GoTaq Probe qPCR Master Mix (Promega Corporation) following the recommended cycling settings suggested by the manufacturer. Reactions were carried out in a volume of 10 μl, including 1 μl of template cDNA. RT-qPCR was conducted in HardShell 384-Well Plates (Bio-Rad). Microseal ‘B’ PCR Plate Sealing Film (Bio-Rad) was used to seal the plates. CFX96 Touch Real-Time PCR Detection System (Bio-Rad) was used to run the TaqMan Gene Expression Assays. To determine relative gene expression levels, the comparative CT method was employed, using mean CT values of NKG7, compared to 18S (for TaqMan Gene Expression Assays).

Statistical analysis

Statistical analysis was performed using GraphPad Prism 6 (GraphPad Software, Boston, MA) and JMP Pro (v19.1.2, SAS Institute, Cary, NC). Analysis of parasite area under the curve (AUC) in mouse experiments was performed using a Mann-Whitney test, while differences in mouse flow cytometry data was tested using a two-way ANOVA with Sidak’s multiple comparisons test or one-way ANOVA with Tukey’s multiple comparisons test, as appropriate. For analysis of PBT-II co-transfer experiments (S2 Fig), a random effects regression model was used to account for the paired measures for each mouse. The model includes Day (post-infection), Condition (PbT-II genotype), and the Day by condition interaction as factors. If the interaction was significant, we then compared the results of the 2 conditions at each day, while accounting for the paired nature of the data. In other mouse experiments, differences in PbTII cell frequencies were assessed using a Wilcoxon test. Analysis of human cell clusters was performed using Edge R, while differences in immune molecule expression by clusters was assessed by one-way ANOVA with Tukey’s multiple comparisons test (Table 8).

Download:

• PNG
  larger image
• TIFF
  original image

Table 8. Resources and equipment [111].

https://doi.org/10.1371/journal.ppat.1013528.t008