Malaria is a blood-borne disease of major global importance that is caused by infection with parasitic protozoa from the genus Plasmodium and afflicts hundreds of millions of people annually (World Health Organization, 2021). Malarial illness presents with a wide array of clinical signs and symptoms that range from minimal to severe and life-threatening (reviewed in Miller et al., 2013; Milner, 2018). Parasite species and genotype, variant antigen expression, infected red blood cell (iRBC) cytoadhesion and sequestration in the microvasculature, iRBC concealment, and divergent host genetics and host immune responses contribute to the range of disease severity observed during Plasmodium infections (Miller et al., 2002; Baird, 2013; Miller et al., 2013; Wassmer et al., 2015; Fonseca et al., 2017). Given the complexities of malaria, the continued rise in drug resistance, and the general need for new parasite and possible host-directed antimalarial treatments, new insights into host-parasite interactions and mechanisms that govern malarial disease progression and resilience are of great importance (Miller et al., 2013; Ashley et al., 2014; Amato et al., 2018).

Plasmodium knowlesi is a zoonotic simian malaria parasite that causes a spectrum of malarial disease manifestations in humans in Southeast Asia (Singh et al., 2004; Cox-Singh et al., 2008; Cox-Singh and Singh, 2008; Daneshvar et al., 2009; William et al., 2011; Barber et al., 2017; Barber et al., 2018; Rajahram et al., 2019). This parasite naturally infects kra monkeys (Macaca fascicularis) and other macaque species that are native to Southeast Asia, and these animals control their parasitemia without treatment, leading to chronic infections (Knowles and Gupta, 1932; Garnham, 1966; Collins et al., 1992). In contrast, if left untreated with antimalarial drugs, P. knowlesi causes severe, lethal disease nearly universally in Indian rhesus macaques (Macaca mulatta) (Knowles and Gupta, 1932; Garnham, 1966; Coatney et al., 1971; Spangler et al., 1978; Coatney et al., 2003). Relevantly, key differences between the transcriptomic responses of M. fascicularis and M. mulatta to P. knowlesi infection were recently documented by Gupta and colleagues, bringing emphasis to the utility of cross-species comparisons (Gupta et al., 2021; Gupta et al., 2022).

Severe P. knowlesi infections in humans have been specifically associated with gastrointestinal (GI) pathology (Cox-Singh et al., 2008; Barber et al., 2013). Autopsy studies from fatal P. knowlesi human cases to better understand this phenomenon and other disease factors are scarce due to cultural considerations (Cox-Singh et al., 2010). Autopsy reports from P. falciparum cases are more common (Menezes et al., 2012), and P. falciparum trophozoite and schizont (T/S)-iRBCs have been documented in the microvasculature of the GI tract, with some studies demonstrating associated histopathology and GI symptoms (Romero et al., 1993; Seydel et al., 2006). Studies with rhesus macaques have reported P. knowlesi iRBCs in the microvasculature of the GI tract and other tissues, suggestive of cytoadhesion and sequestration, despite – unlike P. falciparum – the peripheral circulation of all asexual blood-stage developmental forms from the life cycle of this species (Miller et al., 1971b; Spangler et al., 1978). However, potential associations between the P. knowlesi parasite burden in the microvasculature of various tissues, adhesive interactions, and tissue damage have yet to be explored.

While all P. knowlesi blood-stage forms circulate, a subpopulation of P. knowlesi iRBCs may cytoadhere and sequester in tissues due to receptor-ligand interactions mediated by schizont-infected cell agglutination (SICA) proteins, which are variant antigens expressed by P. knowlesi that are related to and share features with P. falciparum erythrocyte membrane protein 1 (PfEMP1) and the genes that encode them (Korir and Galinski, 2006; Bunnik et al., 2019). SICA and PfEMP1 are large multi-domain proteins (~200-300 kDa) (Howard et al., 1983; Leech et al., 1984) that are encoded by large, diverse multigene families termed SICAvar and var, respectively (Baruch et al., 1995; Smith et al., 1995; Su et al., 1995; Al-Khedery et al., 1999). The expressed SICA proteins become inserted in the iRBC membrane with variant domains exposed at the surface of SICA positive (SICA[+]) iRBCs (Howard et al., 1983), and their expression in rhesus monkeys has been associated with virulence (Barnwell et al., 1983b). SICAvar gene and protein expression are dependent upon the presence of the spleen, with SICA[+] parasites taking on a SICA negative (SICA[-]) phenotype after passage in splenectomized rhesus (Barnwell et al., 1982; Barnwell et al., 1983a). The SICA[-] phenotype is characterized by downregulated SICAvar and SICA protein expression, notably to an ‘off state’ (Lapp et al., 2013). Infection of spleen-intact rhesus monkeys with SICA[-] parasites may result in milder disease, or return of SICA expression along with severe outcomes and fatality (Barnwell et al., 1983a). Similarly in in vitro cultures, SICAvar gene family expression was shown to be downregulated (Lapp et al., 2015). Plasmodium falciparum variant antigen expression has also been shown to be spleen-dependent in vivo in nonhuman primate (NHP) models and human infections (David et al., 1983; Hommel et al., 1983; Bachmann et al., 2009). SICA protein expression may mediate cytoadhesion and sequestration of a subpopulation of P. knowlesi iRBCs in the microvasculature of tissues, akin to P. falciparum T/S-iRBCs populations that express PfEMP-1 (Aley et al., 1984; Leech et al., 1984; Biggs et al., 1991; Biggs et al., 1992), enabling widespread host-parasite receptor-ligand interactions and sequestration of P. falciparum iRBCs in the tissues (reviewed in Smith, 2014). Whether and to what degree P. knowlesi parasite sequestration concomitant with local and systemic inflammation may associate with enhanced symptomology and histopathology, which is evident with PfEMP-1 expression, cytoadhesion, and sequestration (reviewed in Craig et al., 2012; Smith et al., 2013; Wahlgren et al., 2017; Lee et al., 2019), are questions of prime interest.

The experimental P. knowlesi-rhesus macaque infection model has provided significant insights into malarial immunity, notably, the major discovery of malarial antigenic variation during the course of longitudinal infections, which entails changes in the antigenic phenotype of iRBCs in concert with the development of antigen-specific immunity (Brown and Brown, 1965). This model then enabled the first identification and characterization of Plasmodium-encoded iRBC surface-exposed immunodominant variant proteins, namely the P. knowlesi SICA proteins (Barnwell et al., 1983a; Howard et al., 1983; Howard and Barnwell, 1984a; Howard and Barnwell, 1984b; Howard and Barnwell, 1985), which preceded the identification of PfEMP1 based on similar extraction and characterization procedures (Howard et al., 1984; Leech et al., 1984; Howard et al., 1988).

The longitudinal P. knowlesi sporozoite-induced infections in malaria-naïve rhesus monkeys presented here delve into histopathology questions pertinent to understanding P. knowlesi infections in humans. These experiments were generated by the Malaria Host-Parasite Interaction Center (MaHPIC) systems biology research consortium, which previously reported the utilization and integration of multi-omic technologies to investigate and compare host-pathogen interactions during acute, chronic, and relapse phases of malarial infections caused by Plasmodium cynomolgi or Plasmodium coatneyi (Joyner et al., 2016; Joyner et al., 2017; Tang et al., 2017; Cordy et al., 2019; Joyner et al., 2019; Tang et al., 2019; Debarry et al., 2020), and P. knowlesi in kra and rhesus monkeys (Gupta et al., 2021; Peterson et al., 2021; Gupta et al., 2022). Within the rhesus monkey cohorts presented here, longitudinal clinical data and pathological relationships were documented and modest relationships were identified between organ pathology and parasite tissue load assessed in 22 tissue types. Of note, extensive margination of SICA[+] iRBCs was observed along the vascular endothelium of the GI tract. Strikingly, SICA[+] iRBC margination was observed with abundant juxtaposition of iRBCs with the endothelium, but this was not observed in the tissues of splenectomized rhesus infected with SICA[-] parasites. In summary, the rhesus macaque animal model was utilized to systematically characterize clinical features of malarial disease, P. knowlesi sequestration and tissue pathology, and the results point to SICA proteins functioning at the interface of iRBCs and the GI tract.

The experimental cohorts from this study (E06, E33, E34, and E35), including the number of animals, sporozoite inoculations, monitoring, treatments, etc., are detailed in Supplemental Table 1 . Schematics with parasitemias and sampling points for each animal by cohort are included in Supplemental Figures 1–4.

This study determined clinical features of infection and host-parasite interactions that underpin disease severity in P. knowlesi infections in rhesus macaques, providing insights into mechanisms of severe illness in human Plasmodium infections. In the absence of treatment, overwhelming parasitemia is the main cause of death for rhesus macaques infected with P. knowlesi (reviewed in Howard, 1984). Importantly, rhesus monkeys in this study became febrile by day 8 post-inoculation of sporozoites, one day on average later than kra monkeys (Peterson et al., 2021), which, as expected, showed natural resilience to experimental P. knowlesi infections without treatment (Knowles and Gupta, 1932; Garnham, 1966; Collins et al., 1992). This ‘one-day’ difference in fever response was apparent even though both macaque species developed blood-stage parasitemias within a typical time frame of 6-7 days from the time of sporozoite inoculation. The later febrile episodes in the rhesus monkeys suggest that they were less able or unable to detect and mount appropriate, timely systemic inhibitory responses to control the blood-stage parasites. Since P. knowlesi-iRBCs multiply approximately 10-fold every 24 hours, this seemingly unremarkable ‘one-day delay’ in rhesus monkey responsiveness is significant. The delayed containment of parasitemia resulted in apparent prolonged inflammation in the rhesus compared to recovery processes in the kra monkeys (Gupta et al., 2021; Gupta et al., 2022).

Bilirubinemia has been associated with severe P. knowlesi malaria in humans (Barber et al., 2013), and it was mildly elevated in the rhesus monkeys in this study prior to them controlling their parasitemia. Fractionated bilirubin measurements were unavailable, however, and given that there was no histological evidence for cholestasis or other hepatic injury, the elevation in total bilirubin may reflect increased hemolysis. The anemia in the rhesus monkeys was mostly mild and may correspond to the relatively modest increase in serum bilirubin observed. Neither AST nor ALT was significantly different from baseline, however an increase in AST was observed. ALT and AST are both measures of cell integrity, with ALT being more specific to the liver. An elevation of AST without a concurrent elevation of ALT in concert with the absence of hepatocyte necrosis is consistent with a picture of nonspecific tissue stress caused by an infection (Giannini et al., 2005). Despite notable histopathology in the kidneys, both prior to and after parasitemic control, no significant elevation in creatine was observed until after control of the infection.

After subcurative antimalarial medication was administered to the acutely infected rhesus, to ensure their survival and to establish chronic infections, the observed infection-based decreased hemoglobin and platelet levels rebounded rapidly in the animals. The relatively mild anemia and thrombocytopenia occurring in the animals are consistent with expectations based on human cases after antimalarial treatment (Daneshvar et al., 2009). Insufficient erythropoiesis has been reported in rhesus macaques infected with P. cynomolgi and P. coatneyi, as well in rodents and humans infected with Plasmodium parasites (Chang et al., 2004; Chang and Stevenson, 2004; Perkins et al., 2011; Moreno et al., 2013; Thawani et al., 2014; Joyner et al., 2016; Tang et al., 2017; Cordy et al., 2019). Likewise, here, the BM response was blunted in the face of P. knowlesi infection, until treatment was provided. Interestingly, in stark contrast, the BM response was timely in kra monkeys infected with P. knowlesi, with the kra monkey BM responding naturally to the infection without treatment (Peterson et al., 2021). This is consistent with their faster overall recovery from infection compared to the rhesus, as indicated by transcriptomics (Gupta et al., 2021; Peterson et al., 2021; Gupta et al., 2022). Future determinations of the mechanism(s) by which hemoglobin levels in rhesus monkeys stabilize after subcurative treatment may lead to a better understanding of malarial anemia in relation to patient care.

Much remains to be learned about the specific systemic host-parasite interactions and biological, immunological, and metabolic cascades that influence the ultimate pathology and disease course of malaria in both humans and NHPs. Tissue data obtained from autopsy and necropsy procedures, respectively, to better understand malaria pathology in the context of clinical data and the clinical spectrum of the disease, can be powerful, providing a holistic assessment far beyond what can be achieved with either cross-sectional samples or multiple time points during infections. Previous studies have reported the presence of P. knowlesi iRBCs in various tissues from infected rhesus monkeys (Miller et al., 1971b; Anderios et al., 2010). The data presented here expand upon that knowledge by quantitatively determining that P. knowlesi iRBCs are preferentially found in specific organs like the spleen, adrenal gland, lung, kidney, and the organs of the GI tract. Similar to prior studies with P. vivax infections of S. boliviensis monkeys (Peterson et al., 2019), a modest relationship was identified here between the rhesus histopathology scores and P. knowlesi iRBC accumulation in these tissues. While P. knowlesi iRBC accumulation in the tissues may directly or indirectly cause histopathology, our data suggest that it is neither necessary nor sufficient. Instead, host responses are likely the main contributors to the observed tissue damage. This hypothesis, which warrants further exploration, is in fact consistent with tissue-damage studies that have directly shown that P. falciparum-iRBC accumulation in the brain is not necessarily the main driver of cerebral malaria (Seydel et al., 2006; Milner et al., 2014; Milner et al., 2015; Coban et al., 2018).

Acute Plasmodium infections from multiple species – whether in NHPs or humans – are known to cause damage to tissues and vital organs, which can contribute to clinical complications and death (Spitz, 1946; Spangler et al., 1978; Cox-Singh et al., 2010; Lacerda et al., 2012; Milner et al., 2014). Therefore, it was anticipated in this study that escalating parasitemias in rhesus would cause tissue damage that, at least in part, would contribute to the inability of this species to survive infection without treatment (Spangler et al., 1978). Surprisingly, the histopathology observations during the acute phase of the rhesus infections proved to be comparable to those reported previously for P. knowlesi-infected kra monkey tissues (Peterson et al., 2021) and dramatic pathological findings in vital organs (e.g., lung, liver, kidney) were not identified. However, gastritis was specifically noted in the rhesus stomach tissue samples acquired from necropsy performed at the time of the acute infections. This finding is noteworthy from a clinical perspective because gastritis and other GI complications and clinical manifestations occur frequently in humans with P. knowlesi malaria (Cox-Singh et al., 2008; Barber et al., 2013).

Striking observations presented here include microscopic evidence for sequestration of P. knowlesi SICA[+] iRBCs with presumptive cytoadhesion that was pronounced in the spleen-intact rhesus. Frank margination of the iRBCs was observed in the mucosal and submucosal vessels of the stomach, duodenum, jejunum, and colon tissues from these animals, and this phenotype was dramatic even with relatively low parasitemias (1-3%) at the time of necropsy. Rheological changes and ‘sludging,’ as reported previously in rhesus infection studies, in part explain capillary congestion with the P. knowlesi-iRBCs (Knisely et al., 1945; Knisely et al., 1964; Miller et al., 1971a; Barber et al., 2018). Yet, the clear predilection shown here for iRBCs at the margins of venules over the lumen implicates iRBC endothelial interaction(s) and cytoadhesion as an apparent mechanism.

A few studies have examined cytoadhesion of P. knowlesi T/S-iRBCs using ex vivo (Fatih et al., 2012) and in vitro (Lee et al., 2022) cultivated parasites, leaving questions as to the parasite ligand(s) involved and the number and types of possible host receptor specificities. Critically, in vitro cultivated P. knowlesi iRBCs have been shown to lose expression of the SICAvar gene family, and these were classified in the ‘off state’ with regards to SICAvar expression (Lapp et al., 2015). Expression of the SICAvar gene family and the encoded SICA proteins were shown to be downregulated to the ‘off state’ in iRBCs from splenectomized hosts, bringing emphasis to the importance of the in vivo splenic environment for regulating SICA expression (Lapp et al., 2013) and ultimately virulence of P. knowlesi infections (Barnwell et al., 1983b). In turn, experimental systems to test for receptor-ligand interactions may require ex vivo parasites obtained from spleen-intact animals. Our preliminary findings in this area (unpublished data), as others (Fatih et al., 2012; Lee et al., 2022), suggest levels of complexity on par with P. falciparum. Therefore, sorting out the precise receptors, ligands, and interactions that support P. knowlesi cytoadhesion and sequestration will not be an easy task.

SICA proteins have been considered prime candidates mediating possible adhesive interaction(s), due to their commonalities with PfEMP1, including iRBC surface localization and multiple diverse externalized cysteine-rich domains (Al-Khedery et al., 1999; Korir and Galinski, 2006); these features are hallmarks of PfEMP1 that account for P. falciparum’s selective T/S-iRBC cytoadhesion characteristics and sequestration (reviewed in Howard, 1984; Smith, 2014). Here, tissues from splenectomized rhesus monkeys infected with SICA[-] parasites lacked the frank margination phenotype observed in the tissues from spleen-intact monkeys infected with SICA[+] parasites. This observation strongly supports the premise that SICA proteins mediate the adhesion of the iRBCs to the endothelium and provides further impetus for studying the SICA[+] and SICA[-] parasites in parallel infections to better understand the pathological mechanisms and biological implications of iRBC adherence to the endothelium. SICA protein expression at the surface of P. knowlesi-iRBCs is associated with virulence in rhesus and SICAvar gene family expression is somehow regulated by the spleen (Barnwell et al., 1983a; Lapp et al., 2013), as is PfEMP1 variant antigen gene and protein expression in Saimiri monkeys and humans (David et al., 1983; Hommel et al., 1983; Bachmann et al., 2009).

Nonetheless, as for PfEMP1, it is worth noting that the biological function of SICA in the context of the iRBC membrane remains unknown (reviewed in Howard, 1984; Galinski et al., 2018). In this context, it is important to recognize that cytoadhesion – as proposed to avoid circulation through the spleen – is not necessarily (nor likely) the sole function of either of these proteins (reviewed in Howard, 1984). It would be a breakthrough if an essential iRBC function were to be discovered for the PfEMP-1 and SICA proteins, commensurate with these parasite species’ evolved strategies to maintain numerous var/SICAvar family members to ensure survival. Clearly, as is known for P. knowlesi and most species of Plasmodium where all blood-stage forms circulate, avoidance of the spleen does not fit as the main function of potentially adhesive surface-exposed proteins. In fact, all other Plasmodium species known to infect humans do not have comparable var/SICAvar genes (P. vivax, P. malariae, P. ovale, P. cynomolgi) (Tachibana et al., 2012; Rutledge et al., 2017; Imwong et al., 2018) and all their iRBC developmental life cycle stages circulate, seemingly unaffected by passage through the spleen (or potentially even benefiting from intra-splenic interactions (reviewed in Kho et al., 2021; Fernandez-Becerra et al., 2022). All iRBC developmental forms of these species circulate despite P. malariae and P. ovale sharing knobby iRBC surface phenotypes (Matsumoto et al., 1986; Li et al., 2010) that are morphologically similar to those in P. falciparum infections (Gruenberg et al., 1983; Aikawa et al., 1996). The iRBCs of each of these species differ from P. knowlesi, which has occasional indentations in the form of caveolae peppering the iRBC and no knobby protrusions (Liu et al., 2019). In fact, P. knowlesi is the simplest of them all, having neither knobs nor caveolae vesicle complexes, which are especially abundant at the iRBC surface of P. vivax- and P. cynomolgi-iRBCs (Aikawa et al., 1975; Akinyi et al., 2012). The simplicity, by comparison, of the P. knowlesi iRBC membrane, even compared to the phylogenetically related ‘knobby, variant and sequestering’ simian malaria parasites P. fragile and P. coatneyi (Garnham, 1966; Fremount and Miller, 1975; Handunnetti et al., 1987; Nakano et al., 1996; Collins et al., 2006; Chien et al., 2016), makes it an attractive parasite model for understanding the essential features of Plasmodium-iRBC parasitism and their evolution.

In sum, this study reports a systematic study of P. knowlesi sporozoite-infected rhesus monkeys that expands our understanding of pathogenic features of severe disease and makes comparisons with comparable data from kra monkeys that support resilience to P. knowlesi malaria (Peterson et al., 2021). Multiple physiological and immunological responses were found to be associated with pathology or resilience. Additionally, an association of SICA variant proteins with cytoadhesion and sequestration came to light in comparisons of SICA[+] and SICA[-] parasites in tissues obtained from spleen-intact and splenectomized rhesus, respectively. Studies using multiple NHP models in biomedical research have been beneficial for several decades towards understanding HIV pathogenesis, specifically with the comparison of SIV-infected rhesus macaques and sooty mangabeys in relation to the development of AIDS (Bosinger et al., 2009; Palesch et al., 2018). Correspondingly, a two-host animal model system that includes rhesus and kra monkeys and longitudinal infections is now poised for significant advances towards understanding host-parasite interactions, cascades and networks. Host mechanistic differences in physiology, immune responses, and pathogenesis, can be unravelled to better understand malaria severity and resilience against P. knowlesi and potentially other species of Plasmodium that can infect these species.

MaHPIC members participating in discussions at the time of the planning, implementation, or analysis of this project include: Dave C. Anderson, Ferhat Ay, Cristiana F. A. Brito, John W. Barnwell, Megan DeBarry, Steven E. Bosinger, Jung-Ting Chien, Jinho Choi, Anuj Gupta, Jay C. Humphrey, Chris Ibegbu, Xuntian Jiang, Dean P. Jones, Nicolas Lackman, Tracey J. Lamb, Frances E.-H. Lee, Karine Gaelle Le Roche, Shuzhao Li, Esmeralda V.S. Meyer, Diego M. Moncada-Giraldo, Dan Ory, Jan Pohl, Saeid Safaei, Ignacio Sanz, Maren Smith, Gregory Tharp, ViLinh Tran, Elizabeth D. Trippe, Karan Uppal, Susanne Warrenfeltz, Tyrone Williams, Zerotti L. Woods.

The clinical datasets generated and analyzed for this study can be found in the NIAID Bioinformatics Resource Center Database, PlasmoDB.org, https://plasmodb.org/plasmo/app/static-content/PlasmoDB/mahpic.html (Aurrecoechea et al., 2009; Amos et al., 2022).

The animal study was reviewed and approved by Emory University’s Institutional Animal Care and Use Committee (IACUC) and the Animal Care and Use Review Office (ACURO) of the US Department of Defense.

Conceived and designed the experiments: MP, CJ, JB, JW, LF, RT, JK, AM, SG, EV, JG, RC, MG, and members of the MaHPIC-Consortium. Performed the experiments: MP, CJ, JB, JW, MC-M, CS, LF, WC, JJ, SL, SS, AH, DM, EK. Performed data analysis: MP, CJ, JB, LF, SG, JG. Interpreted the data analysis: MP, CJ, RC, MG and members of the MaHPIC-Consortium. Managed and led validation and quality control of datasets for clinical and telemetry results and deposited the data and metadata: MN, JD, JK. Generated the figures: MP, CJ, LF, SG. Wrote the paper: MP, MG. Provided manuscript editorial contributions: CJ, SL, LF, JD, RT, JK, AM, SG, EV, JG, RC. All authors reviewed the manuscript. All authors read and approved the final manuscript.

This project was funded in part by the National Institute of Allergy and Infectious Diseases; National Institutes of Health, Department of Health and Human Services, which established the MaHPIC [Contract No. HHSN272201200031C; MG], the NIH Office of Research Infrastructure Programs/OD P51OD011132, the Defense Advanced Research Program Agency and the US Army Research Office via a cooperative agreement [Contract No. W911NF16C0008; MG], which funded the Technologies for Host Resilience-Host Acute Models of Malaria to study Experimental Resilience (THoR’s HAMMER) consortium.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.