Early innate cytokine production

Serum cytokine profiling and transcriptional analysis at defined time points during primary CHMI has consistently shown increased levels of the type II interferon IFNγ shortly after the onset blood-stage infection and prior to development of symptoms (Fig. 2a) (Harpaz et al. Reference Harpaz, Edelman, Wasserman, Levine, Davis and Sztein1992; Hermsen et al. Reference Hermsen, Konijnenberg, Mulder, Loe, van Deuren, van der Meer, van Mierlo, Eling, Hack and Sauerwein2003; Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005; Ockenhouse et al. Reference Ockenhouse, Hu, Kester, Cummings, Stewart, Heppner, Jedlicka, Scott, Wolfe, Vahey and Burke2006; Bijker et al. Reference Bijker, Bastiaens, Teirlinck, van Gemert, Graumans, van de Vegte-Bolmer, Siebelink-Stoter, Arens, Teelen, Nahrendorf, Remarque, Roeffen, Jansens, Zimmerman, Vos, van Schaijk, Wiersma, van der Ven, de Mast, van Lieshout, Verweij, Hermsen, Scholzen and Sauerwein2013; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014; Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014). The IFNγ-induced chemokine CXCL9 has been detected in the vast majority of volunteers examined, and correlated with IFNγ production (Bijker et al. Reference Bijker, Bastiaens, Teirlinck, van Gemert, Graumans, van de Vegte-Bolmer, Siebelink-Stoter, Arens, Teelen, Nahrendorf, Remarque, Roeffen, Jansens, Zimmerman, Vos, van Schaijk, Wiersma, van der Ven, de Mast, van Lieshout, Verweij, Hermsen, Scholzen and Sauerwein2013; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014). In contrast, secretion or transcription of other cytokines including IL-1α, IL-1β, TNFα, IL-6, IL-8, IL-12p40, IL-12p70, TGFβ and IL-10 have been detected only in a fraction of volunteers (Harpaz et al. Reference Harpaz, Edelman, Wasserman, Levine, Davis and Sztein1992; Hermsen et al. Reference Hermsen, Konijnenberg, Mulder, Loe, van Deuren, van der Meer, van Mierlo, Eling, Hack and Sauerwein2003; Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005; Ockenhouse et al. Reference Ockenhouse, Hu, Kester, Cummings, Stewart, Heppner, Jedlicka, Scott, Wolfe, Vahey and Burke2006; De Mast et al. Reference De Mast, Sweep, McCall, Geurts-Moespot, Hermsen, Calandra, Netea, Sauerwein and van der Ven2008; de Mast et al. Reference de Mast, van Dongen-Lases, Swinkels, Nieman, Roestenberg, Druilhe, Arens, Luty, Hermsen, Sauerwein and van der Ven2009b; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014). This suggests heterogeneity in the innate responses to the malaria parasite, in line with inter-individual differences in the ability of naïve individuals to control Plasmodium blood-stage infections, as evidenced in malariotherapy studies conducted in the 1940–1960s (Molineaux et al. Reference Molineaux, Trauble, Collins, Jeffery and Dietz2002). Walther et al. showed that malaria-naïve volunteers can respond to CHMI in two different manners: (i) rapid pro-inflammatory cytokine production in association with more rapid parasite control but linked with more severe clinical symptoms or (ii) early immune-suppressive TGFβ production associated with weaker parasite control but fewer clinical symptoms (Fig. 2b) (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005, Reference Walther, Woodruff, Edele, Jeffries, Tongren, King, Andrews, Bejon, Gilbert, De Souza, Sinden, Hill and Riley2006). In addition, acute blood-stage during primary CHMI has been shown to modulate responsiveness to TLR ligands and increase inflammatory responsiveness to the parasite itself (McCall et al. Reference McCall, Netea, Hermsen, Jansen, Jacobs, Golenbock, van der Ven and Sauerwein2007). In this study, TLR priming during CHMI correlated with fever and levels of the inflammation-induced, liver-derived C-reactive protein (CRP) (McCall et al. Reference McCall, Netea, Hermsen, Jansen, Jacobs, Golenbock, van der Ven and Sauerwein2007). These combined findings suggest that some volunteers appear to be more susceptible in exacerbating their inflammatory response to the parasite.

Fig. 2. Early immune activation during a primary acute blood-stage infection. In the days after emergence from the liver, blood-stage parasite exposure during CHMI results in a sequence of events of immune activation. Pro-inflammatory cytokines are released (a); likely by innate cells including NK-cells, γδT-cells (c) and monocytes. Pro-inflammatory cytokine release can be preceded by release of TGFβ (b), which limits the ensuing pro-inflammatory response. Cytokines such as IFNγ activate monocytes, which release BAFF (e). Elevated BAFF levels are associated with B-cell re-distribution (h). While absolute lymphocyte numbers decrease during acute infection (g), individual B-cell subsets such as plasma blasts, classical and atypical MBC (h), γδT-cells (d) and Tregs (i) proliferate and expand. Treg activity appears to contribute to the reduced proliferative capacity of T-cells during acute infection (j). DCs show transiently enhanced apoptosis and reduced endocytic capacity (f).BAFF, B-cell activating factor; CHMI, controlled human malaria infection; DCs, dendritic cells; MBC, memory B-cells, Tregs, regulatory T-cells

Innate immune cell activation during CHMI

The cellular basis of these differences in innate responses might relate to heterogeneity in distribution of innate immune cell subsets such as monocytes, natural killer (NK)- and γδT-cells. Rapid production of IFNγ during CHMI (Harpaz et al. Reference Harpaz, Edelman, Wasserman, Levine, Davis and Sztein1992; Hermsen et al. Reference Hermsen, Konijnenberg, Mulder, Loe, van Deuren, van der Meer, van Mierlo, Eling, Hack and Sauerwein2003; Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005; Ockenhouse et al. Reference Ockenhouse, Hu, Kester, Cummings, Stewart, Heppner, Jedlicka, Scott, Wolfe, Vahey and Burke2006; Bijker et al. Reference Bijker, Bastiaens, Teirlinck, van Gemert, Graumans, van de Vegte-Bolmer, Siebelink-Stoter, Arens, Teelen, Nahrendorf, Remarque, Roeffen, Jansens, Zimmerman, Vos, van Schaijk, Wiersma, van der Ven, de Mast, van Lieshout, Verweij, Hermsen, Scholzen and Sauerwein2013; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014; Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014) is likely linked to NK- and γδT-cells, which are also the prime producers of IFNγ in response to P. falciparum-infected red blood cells (PfRBCs) in vitro (Fig. 2c) (Korbel et al. Reference Korbel, Newman, Almeida, Davis and Riley2005; D'Ombrain et al. Reference D'Ombrain, Hansen, Simpson and Schofield2007; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011, Reference Teirlinck, Roestenberg, van de Vegte-Bolmer, Scholzen, Heinrichs, Siebelink-Stoter, Graumans, van Gemert, Teelen, Vos, Nganou-Makamdop, Borrmann, Rozier, Erkens, Luty, Hermsen, Sim, van Lieshout, Hoffman, Visser and Sauerwein2013; Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015). Another feature of in vitro PfRBC-stimulated NK cells is de-granulation and granzyme production (Korbel et al. Reference Korbel, Newman, Almeida, Davis and Riley2005). And while a sequential increase in plasma granyzmes A and B has also been shown during CHMI (Hermsen et al. Reference Hermsen, Konijnenberg, Mulder, Loe, van Deuren, van der Meer, van Mierlo, Eling, Hack and Sauerwein2003), the source of these cytotoxic mediatos in vivo remains to be determined.

γδT-cells are activated during blood-stage infection after CHMI, as shown by their sequential up-regulation of the activation markers CD69 and human leukocyte antigen (HLA)-DR, as well as CD45RO, which marks transition to a memory phenotype (Rzepczyk et al. Reference Rzepczyk, Stamatiou, Anderson, Stowers, Cheng, Saul, Allworth, McCormack, Whitby, Olive and Lawrence1996). At least in vitro, NK- and γδT-cells are not activated uniformly; instead activation appears to be linked to individual subsets and is influenced by differences in NK receptor expression or polymorphisms of killer Ig-like receptors (Artavanis-Tsakonas et al. Reference Artavanis-Tsakonas, Eleme, McQueen, Cheng, Parham, Davis and Riley2003; Korbel et al. Reference Korbel, Newman, Almeida, Davis and Riley2005; D'Ombrain et al. Reference D'Ombrain, Hansen, Simpson and Schofield2007; McCall et al. Reference McCall, Roestenberg, Ploemen, Teirlinck, Hopman, de Mast, Dolo, Doumbo, Luty, van der Ven, Hermsen and Sauerwein2010; Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015). Differential activation of individual NK- or γδT-cells subsets in vivo during CHMI, however, has not been examined to date. Equally unexplored is the contribution of so-called innate lymphoid cells, which can direct innate and adaptive responses, to the early inflammatory response during CHMI (Artis and Spits, Reference Artis and Spits2015).

Of note, activation of the innate compartment not only occurs during acute infection, but also appears to extend beyond a primary CHMI, illustrated by sustained expansion of the γδT-cell compartment several weeks after drug cure (Rzepczyk et al. Reference Rzepczyk, Stamatiou, Anderson, Stowers, Cheng, Saul, Allworth, McCormack, Whitby, Olive and Lawrence1996; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011; Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015), and even up to over a year after a single infection (Fig. 2d) (Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011). Such as expansion of the γδT-cell compartment has also been shown after recovery from a single naturally acquired infection in travellers to malaria endemic areas (Martini et al. Reference Martini, Paglia, Montesano, Enders, Gentile, Pauza, Gioia, Colizzi, Narciso, Pucillo and Poccia2003). It remains unclear whether this expansion is polyclonal or restricted to specific γδT-cell subsets. That expansion is not only observed after blood-stage infection induced by inoculation with asexual parasite (Rzepczyk et al. Reference Rzepczyk, Stamatiou, Anderson, Stowers, Cheng, Saul, Allworth, McCormack, Whitby, Olive and Lawrence1996) or fully infectious sporozoites (Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011; Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015), but also after (repeated) injection of irradiated sporozoites (Seder et al. Reference Seder, Chang, Enama, Zephir, Sarwar, Gordon, Holman, James, Billingsley, Gunasekera, Richman, Chakravarty, Manoj, Velmurugan, Li, Ruben, Li, Eappen, Stafford, Plummer, Hendel, Novik, Costner, Mendoza, Saunders, Nason, Richardson, Murphy, Davidson and Richie2013) suggests that both life-cycle stages contribute to γδT-cell expansion. Moreover, both γδT- and NK-cells show memory-like enhanced IFNγ responses upon re-stimulation with PfRBC in vitro at several weeks after parasite clearance (Fig. 3a) (McCall et al. Reference McCall, Roestenberg, Ploemen, Teirlinck, Hopman, de Mast, Dolo, Doumbo, Luty, van der Ven, Hermsen and Sauerwein2010; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011; Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015) At least in NK-cells, this seems to depend on enhanced secretion of T-cell derived cytokines such as IL-2 (Horowitz et al. Reference Horowitz, Newman, Evans, Korbel, Davis and Riley2010; McCall et al. Reference McCall, Roestenberg, Ploemen, Teirlinck, Hopman, de Mast, Dolo, Doumbo, Luty, van der Ven, Hermsen and Sauerwein2010).

Fig. 3. Maintenance of parasite-specific memory responses after a resolved, primary CHMI. Following a single controlled P. falciparum infection, immune memory is induced and maintained in multiple compartments. This includes memory-like enhanced IFNγ production by innate lymphocytes such as NK- and γδT-cells (a). Single- and pluripotent EM T-cells are induced at a greater magnitude and maintained more stable than CM T-cell responses (b). Parasite-specific antibodies and MBC are also readily induced, and IgG responses are more stable than IgM responses (c).CM, central memory; CHMI, controlled human malaria infection; EM, effector memory; IFN, interferon; MBC, memory B-cells.

Monocytes come in different ‘flavours’ (Mitchell et al. Reference Mitchell, Roediger and Weninger2014), and all the three main monocytes subsets become activated after CHMI, as evidenced by enhanced BAFF production. (Fig. 2e) (Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014). Monocyte activation could be mediated directly by the parasite via pattern recognition receptors, induced cytokines such as IFNγ, but also by anaphylotoxins such as C5a resulting from complement activation (Conroy et al. Reference Conroy, Serghides, Finney, Owino, Kumar, Gowda, Liles, Moore and Kain2009), as observed in CHMI (Roestenberg et al. Reference Roestenberg, McCall, Mollnes, van Deuren, Sprong, Klasen, Hermsen, Sauerwein and van der Ven2007). Monocytes are the main TGFβ-producing PBMC subset during CHMI and TGFβ-positive monocytes are particularly increased in those individuals with a more balanced, less inflammatory response (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005). It remains to be determined which monocyte subset is responsible for this regulatory response, and which subset mediates pro-inflammatory priming as proposed previously (McCall et al. Reference McCall, Netea, Hermsen, Jansen, Jacobs, Golenbock, van der Ven and Sauerwein2007).

Monocytes as well as dendritic cells (DCs) are important mediators for both innate and adaptive immune responses. All monocytes subsets as well as blood dendritic cell antigen (BDCA)-1+ DCs increase B-cell activating factor (BAFF) expression during CHMI (Fig. 2e), suggesting that they contribute to directing B-cell activation or homeostasis during P. falciparum infection (Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014). On the other hand, circulating DCs showed increased apoptosis and decreased endocytic activity during early blood-stage infection (Fig. 2f) (Woodberry et al. Reference Woodberry, Minigo, Piera, Amante, Pinzon-Charry, Good, Lopez, Engwerda, McCarthy and Anstey2012). This could indicate impaired DC function as suggested from murine and in vitro models and field work (Wykes and Good, Reference Wykes and Good2008). Of note, reduced endocytic activity, at least when determined in vitro, is a typical consequence of DC maturation (Liu and Roche, Reference Liu and Roche2015), and as an important mechanism to prevent immunopathology, DCs increase their susceptible to major histocompatibility complex (MHC) class II and type I interferon-driven apoptosis upon activation and maturation (Bertho et al. Reference Bertho, Blancheteau, Setterblad, Laupeze, Lord, Drenou, Amiot, Charron, Fauchet and Mooney2002; Kushwah and Hu, Reference Kushwah and Hu2010; Fuertes Marraco et al. Reference Fuertes Marraco, Scott, Bouillet, Ives, Masina, Vremec, Jansen, O'Reilly, Schneider, Fasel, Shortman, Strasser and Acha-Orbea2011). Therefore, apoptosis and reduced endocytosis during CHMI might also simply be a consequence of parasite-induced DC maturation. Successful parasite-induced DC activation would also be in line with the successful priming of parasite-specific T-cell responses, as outlined below (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005; Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011, Reference Teirlinck, Roestenberg, van de Vegte-Bolmer, Scholzen, Heinrichs, Siebelink-Stoter, Graumans, van Gemert, Teelen, Vos, Nganou-Makamdop, Borrmann, Rozier, Erkens, Luty, Hermsen, Sim, van Lieshout, Hoffman, Visser and Sauerwein2013; Orlov et al. Reference Orlov, Vaida, Finney, Smith, Talley, Wang, Kappe, Deng, Schooley and Duffy2012; Elias et al. Reference Elias, Collins, Halstead, Choudhary, Bliss, Ewer, Sheehy, Duncan, Biswas, Hill and Draper2013; Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014). The basically unaltered expression of HLA-DR on circulating myeloid DCs as well as the slight reduction on plasmacytoid DCs (Woodberry et al. Reference Woodberry, Minigo, Piera, Amante, Pinzon-Charry, Good, Lopez, Engwerda, McCarthy and Anstey2012) is puzzling in this context. A possible explanation could be that parasite activated DCs might preferentially migrate out of the circulation, as suggested previously based on field and in vitro data (Pichyangkul et al. Reference Pichyangkul, Yongvanitchit, Kum-arb, Hemmi, Akira, Krieg, Heppner, Stewart, Hasegawa, Looareesuwan, Shanks and Miller2004). Taken together, whether Plasmodium infection in humans induces or impairs DC activation and functionality and how this affects T-cell priming remains an important question for future investigations.

Infection-induced lymphocyte sequestration

In parallel with emerging parasitaemia and immune activation, transient lymphopenia has been consistently observed in the early phases of a blood-stage infection (Fig. 2g) (Rzepczyk et al. Reference Rzepczyk, Stamatiou, Anderson, Stowers, Cheng, Saul, Allworth, McCormack, Whitby, Olive and Lawrence1996; Church et al. Reference Church, Le, Bryan, Gordon, Edelman, Fries, Davis, Herrington, Clyde, Shmuklarsky, Schneider, McGovern, Chulay, Ballou and Hoffman1997; De Mast et al. Reference De Mast, Sweep, McCall, Geurts-Moespot, Hermsen, Calandra, Netea, Sauerwein and van der Ven2008; Woodberry et al. Reference Woodberry, Minigo, Piera, Amante, Pinzon-Charry, Good, Lopez, Engwerda, McCarthy and Anstey2012; Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014). A likely explanation is sequestration (Rzepczyk et al. Reference Rzepczyk, Stamatiou, Anderson, Stowers, Cheng, Saul, Allworth, McCormack, Whitby, Olive and Lawrence1996), since lymphocyte subset numbers recover shortly after treatment (Woodberry et al. Reference Woodberry, Minigo, Piera, Amante, Pinzon-Charry, Good, Lopez, Engwerda, McCarthy and Anstey2012). Increased secretion of the chemokine CXCL9 (Ockenhouse et al. Reference Ockenhouse, Hu, Kester, Cummings, Stewart, Heppner, Jedlicka, Scott, Wolfe, Vahey and Burke2006; Bijker et al. Reference Bijker, Bastiaens, Teirlinck, van Gemert, Graumans, van de Vegte-Bolmer, Siebelink-Stoter, Arens, Teelen, Nahrendorf, Remarque, Roeffen, Jansens, Zimmerman, Vos, van Schaijk, Wiersma, van der Ven, de Mast, van Lieshout, Verweij, Hermsen, Scholzen and Sauerwein2013; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014) and a selective reduction in circulating B-cell populations expressing the CXCL9 receptor CXCR3 during acute blood-stage (Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014) corroborate this explanation. Within the B-cell compartment, changes in proportions of different B-cell subsets could further be linked to their degree of expression of the receptor for BAFF (Fig. 2h) (Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014), a cytokine that has previously been shown to alter the responsiveness of B-cells for chemotactic stimuli (Badr et al. Reference Badr, Borhis, Lefevre, Chaoul, Deshayes, Dessirier, Lapree, Tsapis and Richard2008). Whether in other lymphocyte populations, subsets with specific chemokine receptor signatures are equally transiently decreased has not yet been investigated.

CHMI-induced T-cell regulation

Induction of regulatory T-cells (Tregs) is a well-reported consequence of blood-stage infection in field settings and model systems (Scholzen et al. Reference Scholzen, Minigo and Plebanski2010) and also evident during CHMI. Ten days after sporozoite inoculation, when blood-stage parasites are usually detected by microscopy and volunteers treated, expression of the Treg transcription factor Foxp3 was increased, correlating with an increase in circulating bioactive TGFβ (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005). CD4 + CD25hiCD69-Tregs were expanded (Fig. 2i) and mediated suppression of parasite-specific responses during infection (Fig. 2j) (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005). That loss of lymphocyte proliferative capacity could be restored by addition of IL-2 (Rzepczyk et al. Reference Rzepczyk, Stamatiou, Anderson, Stowers, Cheng, Saul, Allworth, McCormack, Whitby, Olive and Lawrence1996) suggests that parasite-induced Tregs might function by competition with effector T-cells for IL-2 (Hofer et al. Reference Hofer, Krichevsky and Altan-Bonnet2012). Further data indicate that this parasite-specific regulatory response can extend beyond acute infection; while 4 weeks after CHMI, in vitro CD4 and CD8 T-cell responses to unrelated antigens were unaffected as such, they were reduced when PfRBC were added to the assay (Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009), indicating a suppressive function of parasite antigen re-called Tregs. The longevity of this regulatory response and life-cycle stage specificity of these Tregs, and therefore their potential impact on subsequent liver- and/or blood-stage infections, remain elusive.

Link between immune activation and parasitaemia

Immune activation and regulation during the early stage of blood-stage infection appear to be dependent directly on duration of exposure and parasite load: after CHMI, peak inflammatory cytokine production correlated with the length of time until a defined level of parasitaemia was reached (Walther et al. Reference Walther, Woodruff, Edele, Jeffries, Tongren, King, Andrews, Bejon, Gilbert, De Souza, Sinden, Hill and Riley2006), and expansion of CD4 + CD25hiCD69-Tregs was the more prominent the earlier parasites reached the threshold of detection (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005). Moreover, peak parasitaemia correlated with subsequent maximum plasma levels of IFNγ and BAFF as well as proliferation of various B-cell subsets including plasma blast, classical and atypical memory B-cells (MBCs) (Fig. 2h) (Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014). While no causal relationships can be directly inferred from such correlations, parasite-driven induction and expansion of Tregs, IFNγ and BAFF production, and B-cell proliferation is consistent with findings from in vitro studies using PfRBC stimulated peripheral blood mononuclear cells (PBMCs) (Donati et al. Reference Donati, Zhang, Chene, Chen, Flick, Nystrom, Wahlgren and Bejarano2004; Scholzen et al. Reference Scholzen, Mittag, Rogerson, Cooke and Plebanski2009; Kumsiri et al. Reference Kumsiri, Potup, Chotivanich, Petmitr, Kalambaheti and Maneerat2010).

Already at the time of first diagnosis, CHMI volunteers show similar pattern recognition receptor and pro-inflammatory cytokine gene transcription as naturally exposed individuals at more advanced stages of infection (Ockenhouse et al. Reference Ockenhouse, Hu, Kester, Cummings, Stewart, Heppner, Jedlicka, Scott, Wolfe, Vahey and Burke2006), while transcription of stress-related heat-shock proteins and other markers of inflammation as well as the induction of (counter)-regulatory pathways are much less pronounced (Ockenhouse et al. Reference Ockenhouse, Hu, Kester, Cummings, Stewart, Heppner, Jedlicka, Scott, Wolfe, Vahey and Burke2006). Markers and signs of inflammation and innate activation then substantially increase in the days after treatment. This includes the frequency of symptoms, fever (Church et al. Reference Church, Le, Bryan, Gordon, Edelman, Fries, Davis, Herrington, Clyde, Shmuklarsky, Schneider, McGovern, Chulay, Ballou and Hoffman1997) and changes in haematological parameters (Church et al. Reference Church, Le, Bryan, Gordon, Edelman, Fries, Davis, Herrington, Clyde, Shmuklarsky, Schneider, McGovern, Chulay, Ballou and Hoffman1997; de Mast et al. Reference de Mast, Groot, Lenting, de Groot, McCall, Sauerwein, Fijnheer and van der Ven2007, Reference De Mast, Sweep, McCall, Geurts-Moespot, Hermsen, Calandra, Netea, Sauerwein and van der Ven2008, Reference de Mast, Nadjm, Reyburn, Kemna, Amos, Laarakkers, Silalye, Verhoef, Sauerwein, Swinkels and van der Ven2009a, Reference de Mast, van Dongen-Lases, Swinkels, Nieman, Roestenberg, Druilhe, Arens, Luty, Hermsen, Sauerwein and van der Venb; Roestenberg et al. Reference Roestenberg, McCall, Mollnes, van Deuren, Sprong, Klasen, Hermsen, Sauerwein and van der Ven2007) as well as production of the acute-phase protein CRP (Harpaz et al. Reference Harpaz, Edelman, Wasserman, Levine, Davis and Sztein1992; Hermsen et al. Reference Hermsen, Konijnenberg, Mulder, Loe, van Deuren, van der Meer, van Mierlo, Eling, Hack and Sauerwein2003), complement activation (Roestenberg et al. Reference Roestenberg, McCall, Mollnes, van Deuren, Sprong, Klasen, Hermsen, Sauerwein and van der Ven2007), cytokine secretion (Hermsen et al. Reference Hermsen, Konijnenberg, Mulder, Loe, van Deuren, van der Meer, van Mierlo, Eling, Hack and Sauerwein2003), B-cell, monocyte and DC activation (Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014). Since this is found across different treatment regimes, the most likely explanation is that the abrupt release of material from drug-killed and otherwise sequestered mature blood-stage parasites into the circulation leads to an initial exacerbation of immune activation, before all parasite material is finally removed.

Priming of parasite-specific T-cell memory

Transcriptional analysis of PBMCs from CHMI volunteers provides circumstantial evidence for the initiation of adaptive immune responses already during the pre-patent period (Ockenhouse et al. Reference Ockenhouse, Hu, Kester, Cummings, Stewart, Heppner, Jedlicka, Scott, Wolfe, Vahey and Burke2006). This includes (i) increased transcription of Fc receptors such as CD16, which can enhance antigen capture (Dobel et al. Reference Dobel, Kunze, Babatz, Trankner, Ludwig, Schmitz, Enk and Schakel2013), (ii) genes involved in antigen processing and presentation such as subunits of the immune proteasome, chaperones involved in MHC class I peptide loading and MHC class II molecules and (iii) glycolytic enzymes, which mark the metabolic switch from naive/resting to effector responses (Yang and Chi, Reference Yang and Chi2012). Consistently, parasite-specific T-cell responses are induced after a single, primary CHMI (Fig. 2k) (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005; Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011, Reference Teirlinck, Roestenberg, van de Vegte-Bolmer, Scholzen, Heinrichs, Siebelink-Stoter, Graumans, van Gemert, Teelen, Vos, Nganou-Makamdop, Borrmann, Rozier, Erkens, Luty, Hermsen, Sim, van Lieshout, Hoffman, Visser and Sauerwein2013; Orlov et al. Reference Orlov, Vaida, Finney, Smith, Talley, Wang, Kappe, Deng, Schooley and Duffy2012; Elias et al. Reference Elias, Collins, Halstead, Choudhary, Bliss, Ewer, Sheehy, Duncan, Biswas, Hill and Draper2013; Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014). Parasite-specific responses are consistently detectable using intact and lysed blood-stage schizonts (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005; Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011, Reference Teirlinck, Roestenberg, van de Vegte-Bolmer, Scholzen, Heinrichs, Siebelink-Stoter, Graumans, van Gemert, Teelen, Vos, Nganou-Makamdop, Borrmann, Rozier, Erkens, Luty, Hermsen, Sim, van Lieshout, Hoffman, Visser and Sauerwein2013; Orlov et al. Reference Orlov, Vaida, Finney, Smith, Talley, Wang, Kappe, Deng, Schooley and Duffy2012; Elias et al. Reference Elias, Collins, Halstead, Choudhary, Bliss, Ewer, Sheehy, Duncan, Biswas, Hill and Draper2013; Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014) or sporozoites as a stimulus (Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011). Responses to individual parasite antigens appear to be variable between CHMI studies: some detect peptide-specific responses to apical membrane protein (AMA)-1 and merozoite surface protein (MSP)-1 (Elias et al. Reference Elias, Collins, Halstead, Choudhary, Bliss, Ewer, Sheehy, Duncan, Biswas, Hill and Draper2013; Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014), while others fail to do for entire panels of sporozoite, liver- and blood-stage antigens (Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011). Relatively low frequency of T-cells specific for individual antigens may explain this discrepancy. Antigen-specific T-cell responses appear largely allele-specific for the challenge strain (Elias et al. Reference Elias, Collins, Halstead, Choudhary, Bliss, Ewer, Sheehy, Duncan, Biswas, Hill and Draper2013), but also show potential for cross-reactivity (Elias et al. Reference Elias, Collins, Halstead, Choudhary, Bliss, Ewer, Sheehy, Duncan, Biswas, Hill and Draper2013; Teirlinck et al. Reference Teirlinck, Roestenberg, van de Vegte-Bolmer, Scholzen, Heinrichs, Siebelink-Stoter, Graumans, van Gemert, Teelen, Vos, Nganou-Makamdop, Borrmann, Rozier, Erkens, Luty, Hermsen, Sim, van Lieshout, Hoffman, Visser and Sauerwein2013).

Memory phenotype and cytokine profile of parasite-specific T-cells

Circulating memory T-cells can broadly be divided into central memory (CM) T-cells that can home to secondary lymphoid organs and largely lack immediate effector function, and effector memory (EM) T-cells with an altered chemokine receptor profile that preferentially home to non-lymphoid tissues (Sallusto et al. Reference Sallusto, Geginat and Lanzavecchia2004). The more recently described tissue resident memory cells do not re-circulate (Schenkel and Masopust, Reference Schenkel and Masopust2014) and can thus not easily be examined in human subjects. Parasite-specific T-cell responses measured in peripheral blood after a single CHMI are dominated by EM T-cells (Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011) and surprisingly stable for over at least a year (Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011). These T-cell responses are largely of Th1 type origin, including secretion of IFNγ, TNFα, IL-2 and MIP-1α (Walther et al. Reference Walther, Tongren, Andrews, Korbel, King, Fletcher, Andersen, Bejon, Thompson, Dunachie, Edele, de Souza, Sinden, Gilbert, Riley and Hill2005; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011, Reference Teirlinck, Roestenberg, van de Vegte-Bolmer, Scholzen, Heinrichs, Siebelink-Stoter, Graumans, van Gemert, Teelen, Vos, Nganou-Makamdop, Borrmann, Rozier, Erkens, Luty, Hermsen, Sim, van Lieshout, Hoffman, Visser and Sauerwein2013; Orlov et al. Reference Orlov, Vaida, Finney, Smith, Talley, Wang, Kappe, Deng, Schooley and Duffy2012), while only few studies were able to detect parasite-specific IL-4 production (Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009; Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014). In contrast to EM T-cells, CM responses are considerably weaker (Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009; Teirlinck et al. Reference Teirlinck, McCall, Roestenberg, Scholzen, Woestenenk, de Mast, van der Ven, Hermsen, Luty and Sauerwein2011), do not correlate in magnitude with EM responses and appear to decline much faster than EM responses (Fig. 3b) (Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009). This weak CM response might simply be due to preferential re-location of CM cells to lymphoid tissues (Brinkman et al. Reference Brinkman, Peske and Engelhard2013). Alternatively, parasite-induced regulatory mechanisms may contribute since high parasite densities associated with lower CM, but not EM, responses at 4 and 12 weeks post-CHMI (Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009). Moreover, the degree of Foxp3 expression during CHMI negatively correlated with IFNγ EM responses as late as 5 months post-CHMI (Todryk et al. Reference Todryk, Walther, Bejon, Hutchings, Thompson, Urban, Porter and Hill2009). Therefore, not just activation and priming, but also counter-regulation of effector T-cell responses is a consequence of a single P. falciparum infection.

The effect of repeated parasitaemia on T-cell memory is less well examined. Cohorts of Dutch and Tanzanian volunteers subjected to CHMI under very similar conditions (Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015) showed a peculiar lack of parasite-specific IFNγ re-call responses in pre-exposed individuals, that failed to increase after CHMI (Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015). Additionally, innate IFNγ responses in pre-exposed Tanzanians were also lower (Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015). Since IFNγ production by innate lymphocytes such as NK-cells is at least partially dependent on T-cell derived cytokines such as IL-2 (Horowitz et al. Reference Horowitz, Newman, Evans, Korbel, Davis and Riley2010; McCall et al. Reference McCall, Roestenberg, Ploemen, Teirlinck, Hopman, de Mast, Dolo, Doumbo, Luty, van der Ven, Hermsen and Sauerwein2010), the reduced responsiveness of adaptive T-cells may be one underlying reason of this lower innate response to the parasite. While no differences in regulatory T-cell levels were found that might also explain these findings, functionality and potential parasite antigen-specific enrichment of Tregs was not examined. Clearly, future studies are needed to investigate whether this difference in both innate and adaptive compartments may be due for instance to skewing to immune signatures other than Th1-type responses. Rather than being a sign of impairment, a more balanced, less Th1 driven cytokine profile may even be beneficial for the human host, since elevated anti-parasite IFNγ responses due to priming in earlier encounters associate with earlier clinical symptoms during blood-stage infection (Bijker et al. Reference Bijker, Bastiaens, Teirlinck, van Gemert, Graumans, van de Vegte-Bolmer, Siebelink-Stoter, Arens, Teelen, Nahrendorf, Remarque, Roeffen, Jansens, Zimmerman, Vos, van Schaijk, Wiersma, van der Ven, de Mast, van Lieshout, Verweij, Hermsen, Scholzen and Sauerwein2013). This would especially be true when in parallel with a less pronounced Th1 cytokine response – as in the above described Tanzanian CHMI cohort (Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015) – other effector mechanisms, such as parasite-specific antibodies that might help to control parasitaemia, are also present.

Generation of antibody and MBC responses

A single CHMI clearly induces production of parasite-specific antibodies, directed against sporozoite and liver-stage antigens as well as the cross-stage antigen MSP-1 (Biswas et al. Reference Biswas, Choudhary, Elias, Miura, Milne, de Cassan, Collins, Halstead, Bliss, Ewer, Osier, Hodgson, Duncan, O'Hara, Long, Hill and Draper2014; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014; Nahrendorf et al. Reference Nahrendorf, Scholzen, Bijker, Teirlinck, Bastiaens, Schats, Hermsen, Visser, Langhorne and Sauerwein2014; Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014; Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015). That a single infection is sufficient to induce parasite-specific antibody responses in a life-cycle exposure dependent manner is consistent with studies of travellers to malaria endemic countries: while recognition of blood-stage antigens or PfRBCs is usually limited to those travellers who experienced a symptomatic or asymptomatic malaria episode (Jelinek et al. Reference Jelinek, Nothdurft and Loscher1995; Cobelens et al. Reference Cobelens, Verhave, Leentvaar-Kuijpers and Kager1998; Seed et al. Reference Seed, Hamzah and Davis2006), seroconversion to circumsporozoite protein (CSP) can also be found in a fraction of those that were protected from blood-stage infection by chemoprophylaxis (Cobelens et al. Reference Cobelens, Verhave, Leentvaar-Kuijpers and Kager1998; Jelinek et al. Reference Jelinek, Bluml, Loscher and Nothdurft1998; Molle et al. Reference Molle, Petersen and Buhl1999; Nothdurft et al. Reference Nothdurft, Jelinek, Bluml, von Sonnenburg and Loscher1999; Knappik et al. Reference Knappik, Peyerl-Hoffmann and Jelinek2002; Belderok et al. Reference Belderok, van den Hoek, Roeffen, Sauerwein and Sonder2013). Also similar to a first naturally acquired infection in previously naive travellers (Elliott et al. Reference Elliott, Payne, Duffy, Byrne, Tham, Rogerson, Brown and Eisen2007), CHMI generates antibody responses to multiple P. falciparum erythrocyte membrane protein-1 alleles (Turner et al. Reference Turner, Wang, Lavstsen, Mwakalinga, Sauerwein, Hermsen and Theander2011), likely because multiple var genes can be transcribed simultaneously (Peters et al. Reference Peters, Fowler, Gatton, Chen, Saul and Cheng2002; Lavstsen et al. Reference Lavstsen, Magistrado, Hermsen, Salanti, Jensen, Sauerwein, Hviid, Theander and Staalsoe2005; Wang et al. Reference Wang, Hermsen, Sauerwein, Arnot, Theander and Lavstsen2009). In addition, antibody maturation occurs: initially, both MSP-1₁₉-specific IgM and IgG titers increase, followed by a quick decline in IgM levels, while IgG titres are maintained with increasing avidity (Fig. 3c) (Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014). B-cell isotype switching is directed by follicular helper T-cells (McHeyzer-Williams et al. Reference McHeyzer-Williams, Okitsu, Wang and McHeyzer-Williams2012), which although mainly confined to lymphoid tissue follicles, can also appear in the circulation (Locci et al. Reference Locci, Havenar-Daughton, Landais, Wu, Kroenke, Arlehamn, Su, Cubas, Davis, Sette, Haddad, Poignard and Crotty2013; Boswell et al. Reference Boswell, Paris, Boritz, Ambrozak, Yamamoto, Darko, Wloka, Wheatley, Narpala, McDermott, Roederer, Haubrich, Connors, Ake, Douek, Kim, Petrovas and Koup2014). The first attempt to associate CHMI-induced T- and B-cell responses showed an inverse correlation of MSP-1 IgG responses with parasite-specific T-cell re-call proliferation and IFNγ production, but the exact phenotype of these cells and typical follicular helper T-cell cytokine production was not investigated (Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014). Clearly, the induction and role of T-cell help for humoral responses in malaria requires further investigation. Next to antibodies, specific MBC responses to CSP and MSP-1 can be directly measured in PBMCs (Fig. 3c) (Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014; Nahrendorf et al. Reference Nahrendorf, Scholzen, Bijker, Teirlinck, Bastiaens, Schats, Hermsen, Visser, Langhorne and Sauerwein2014). The magnitude of MSP-1₁₉ specific antibody and MBC responses after a primary CHMI correlates with the degree of parasite exposure, determined by both duration and magnitude of blood-stage infection (Biswas et al. Reference Biswas, Choudhary, Elias, Miura, Milne, de Cassan, Collins, Halstead, Bliss, Ewer, Osier, Hodgson, Duncan, O'Hara, Long, Hill and Draper2014; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014; Walker et al. Reference Walker, Okitsu, Porter, Duncan, Amacker, Pluschke, Cavanagh, Hill and Todryk2014). This exposure-dependency of humoral responses is also consistent with similar findings after repeated CHMI (Nahrendorf et al. Reference Nahrendorf, Scholzen, Bijker, Teirlinck, Bastiaens, Schats, Hermsen, Visser, Langhorne and Sauerwein2014).

Effect of blood-stage exposure on maintenance of humoral immune responses

There is a general notion that blood-stage parasites deregulate B-cell function, with negative effects on the maintenance of B-cell memory and antibody responses (Portugal et al. Reference Portugal, Pierce and Crompton2013; Scholzen and Sauerwein, Reference Scholzen and Sauerwein2013). However, CHMI data indicate that blood-stage parasite exposure has per se no negative impact on B-cell memory: despite transient loss of B-cells from the circulation during CHMI-induced acute blood-stage infection (Rzepczyk et al. Reference Rzepczyk, Stamatiou, Anderson, Stowers, Cheng, Saul, Allworth, McCormack, Whitby, Olive and Lawrence1996; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014; Scholzen et al. Reference Scholzen, Teirlinck, Bijker, Roestenberg, Hermsen, Hoffman and Sauerwein2014), previously vaccine- or whole sporozoite-induced antibody and MBC responses to liver- and cross-stage antigens were maintained rather than reduced after CHMI (Biswas et al. Reference Biswas, Choudhary, Elias, Miura, Milne, de Cassan, Collins, Halstead, Bliss, Ewer, Osier, Hodgson, Duncan, O'Hara, Long, Hill and Draper2014; Elias et al. Reference Elias, Choudhary, de Cassan, Biswas, Collins, Halstead, Bliss, Ewer, Hodgson, Duncan, Hill and Draper2014; Nahrendorf et al. Reference Nahrendorf, Scholzen, Bijker, Teirlinck, Bastiaens, Schats, Hermsen, Visser, Langhorne and Sauerwein2014). Blood-stage exposure during CHMI has further no negative impact on specific antibody avidity or functionality (Biswas et al. Reference Biswas, Choudhary, Elias, Miura, Milne, de Cassan, Collins, Halstead, Bliss, Ewer, Osier, Hodgson, Duncan, O'Hara, Long, Hill and Draper2014), and can even boost previous experimentally or naturally induced responses by 2–10-fold (Biswas et al. Reference Biswas, Choudhary, Elias, Miura, Milne, de Cassan, Collins, Halstead, Bliss, Ewer, Osier, Hodgson, Duncan, O'Hara, Long, Hill and Draper2014; Nahrendorf et al. Reference Nahrendorf, Scholzen, Bijker, Teirlinck, Bastiaens, Schats, Hermsen, Visser, Langhorne and Sauerwein2014; Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015). Of course, due to rapid treatment upon diagnosis, maximum densities of parasitaemia after CHMI range from only 1 to 100 parasites/μL⁻¹ (Roestenberg et al. Reference Roestenberg, O'Hara, Duncan, Epstein, Edwards, Scholzen, van der Ven, Hermsen, Hill and Sauerwein2012), i.e. much lower than in natural infections. Nevertheless, even a history of naturally acquired infections does not impair B-cell memory re-called by CHMI; pre-existing antibody responses in Tanzanian volunteers were readily detectable in a large proportion of volunteers and, compared with malaria-naïve Dutch individuals, increased more strongly upon CHMI. This occurred even in sero-negative Tanzanians, providing evidence for a robust MBC response that was stably maintained even in those individuals in which plasma blast-produced antibody levels dropped below detection (Obiero et al. Reference Obiero, Shekalaghe, Hermsen, Mpina, Bijker, Roestenberg, Teelen, Billingsley, Sim, James, Daubenberger, Hoffman, Abdulla, Sauerwein and Scholzen2015). Similarly, antibody responses were also boosted in P. vivax-challenged Colombian individuals (Arevalo-Herrera et al. Reference Arevalo-Herrera, Forero-Pena, Rubiano, Gomez-Hincapie, Martinez, Lopez-Perez, Castellanos, Cespedes, Palacios, Onate and Herrera2014). Together, these CHMI data support that the notion that slow development of clinically protective B-cell responses to malaria might not relate to induction or maintenance of memory itself but rather the polymorphic nature of malarial antigens in the wide variety of genetically distinct field strains (Struik and Riley, Reference Struik and Riley2004).

Concluding remarks

Taken together, CHMI trials have revealed that the early phase of a primary P. falciparum infection is characterized by immune cell activation, re-distribution and inflammatory cytokine production, which coincide with blood-stage infection and are related to the degree of parasitaemia. Heterogeneity in the early inflammatory response appears to relate to differential presentation of clinical symptoms. Moreover, parasite-specific T- and B-cell memory responses are readily induced by a single infection and maintained for prolonged periods of time in the absence of re-exposure. B-cell responses are not perturbed by low dose blood-stage re-exposure, while parasitaemia does show some regulatory influence on Th1 type T-cell memory, which might contribute to limiting inflammation in ongoing and future encounters. Based on these findings, questions arise regarding the role of individual cell subsets in the early inflammatory/regulatory response to the parasite, the induction of adaptive responses and the quality of the immune response in future infections (Box 2). An exciting new development in CHMI is the controlled re-exposure of individuals with a history of natural pre-exposure. Comparative analysis of malaria-naive and naturally exposed individuals has already provided some insights into the effect of pre-exposure on parasite-specific re-call responses. It is further a promising approach to dissect differences in innate and adaptive immune cell activation during acute infection, and may provide novel insights into malaria-associated immune modulation leading to altered disease susceptibility.