The intersection of host in vivo metabolism and immune responses to infection with kinetoplastid and apicomplexan parasites

SUMMARY

Protozoan parasite infection dramatically alters host metabolism, driven by immunological demand and parasite manipulation strategies. Immunometabolic checkpoints are often exploited by kinetoplastid and protozoan parasites to establish chronic infection, which can significantly impair host metabolic homeostasis. The recent growth of tools to analyze metabolism is expanding our understanding of these questions. Here, we review and contrast host metabolic alterations that occur in vivo during infection with Leishmania, trypanosomes, Toxoplasma, Plasmodium, and Cryptosporidium. Although genetically divergent, there are commonalities among these pathogens in terms of metabolic needs, induction of the type I immune responses required for clearance, and the potential for sustained host metabolic dysbiosis. Comparing these pathogens provides an opportunity to explore how transmission strategy, nutritional demand, and host cell and tissue tropism drive similarities and unique aspects in host response and infection outcome and to design new strategies to treat disease.

INTRODUCTION

Parasite infection is a comparatively understudied but potentially powerful system to understand the metabolic regulation of the immune response. Kinetoplastids and apicomplexans are phylogenetically related classes of unicellular protozoan parasites with genus-specific transmission cycles and species-specific euthermic host tropisms. As eukaryotic cells, these parasites have complex metabolic demands that are often much more similar to their mammalian host organisms than bacterial pathogens. They have long evolutionary relationships with their hosts, with many strategies to subvert immune clearance and manipulate host metabolite synthesis, signaling, and trafficking (1–3). A picture is emerging where many of these parasites infect and exploit tissues central to metabolism as part of their pathogenesis strategy. Moreover, many of them have strategies for persistent infection [e.g., references (4, 5)], which poses a unique model to interrogate chronic inflammatory and metabolic pathways and can be used to understand chronic metabolic diseases without relying on genetic tools which can lack precision and finesse.

As tools to reliably measure and quantify metabolites have become increasingly available, there has been a renaissance in the field of immuno-metabolism. This has triggered a revision of historical models of immune response development, and there are many detailed reviews on this topic (6, 7). Microbial components are detected by pathogen sensors (e.g., toll-like receptors, RNA helicases) that initiate inflammatory signaling. To fuel the production of antimicrobial effectors and immune signaling molecules, innate immune signaling is coupled with glycolytic metabolism and activation of the pentose phosphate pathway to synthesize ATP (8). This catabolic response is associated with proinflammatory cytokines including IL-1, IL-12, tumor necrosis factor (TNF)-α, and Interferon (IFN)-γ. This inflammatory signature is often referred to as “classical activation” in antigen-presenting cells (macrophage and dendritic cells), leading to Type 1T helper (TH1) cell response that culminates in CD8 effector T cell activation. In contrast, interaction with extracellular parasites, allergens, and wound healing environments that elicit IL-4, IL-13, and transforming growth factor (TGF-ß) lead to alternative activation of antigen-presenting cells, often associated with eosinophil and mast cell activation, resulting in a type 2T helper (TH2) response that relies more heavily on humoral, B cell immunity to drive host defense. These pathways divert glucose, fatty acids, and glutamine to oxidative phosphorylation (8). TH17 is elicited in response to IL-17 and IL-23 particularly in the intestine and central nervous system. TH1 and TH17 immune responses are pro-inflammatory, catabolic, and come at the potential cost of damage to self. These responses have co-evolved with a second wave of anti-inflammatory responses, the hallmark of which are IL-10 secretion and regulatory T cell responses that transition acute inflammation to a pro-resolution, wound healing TH2 profile (9). Moreover, sensitive tissues like the lung and lipid-rich tissues like adipose depots are predisposed toward a TH2 environment (10). Protozoan parasites can exploit both inflammatory and metabolic arms of this TH1/TH2 ‘toggle switch’ to facilitate long-term persistent infection (11). In keeping with this, chronic infection with these parasites can lead to chronic metabolic dysbiosis, including stunting in children and chronic muscle wasting in adults (7, 12, 13).

Here, we review the host metabolic changes that occur in vivo during infection with the kinetoplastids Leishmania, Trypanosome cruzi, Trypanosome brucei as well as the apicomplexans Toxoplasma gondii, Plasmodium sp., and Cryptosporidium. This cross-system comparison is designed to shed light on similarities in metabolic regulation that may be related to transmission strategy, niche colonization, immune polarization, and/or metabolic demands of the parasite.

KINETOPLASTID PARASITES

Kinetoplastids are flagellated protozoan parasites possessing a DNA-containing structure in their mitochondria located near the flagellum. The kinetoplastids include Trypanosoma and Leishmania parasites causing human illnesses associated with robust immune and metabolic responses. During infection, host metabolic responses can result in promotion or limitation of parasite growth and persistence (1). These metabolic alterations serve as a convergence point that can define disease phenotype caused by kinetoplastid parasitic infection.

Leishmania

Leishmania infect humans through the bite of infected female sandflies, which introduce Leishmania promastigotes into the skin. Macrophages, neutrophils, and dendritic cells phagocytose promastigotes, forming a parasitophorous vacuole in which parasites transform to amastigotes and proliferate. Amastigotes subsequently lyse the host cell to infect other phagocytes in the host. Leishmania infection can result in three different clinical forms of leishmaniasis: cutaneous (CL), mucocutaneous (MCL), and visceral leishmaniasis (VL). This is determined by the species and strain of the parasite as well as host immune polarization (14). L. major infection was a model system used to understand the importance of TH1 vs TH2 immunity: C57BL/6 mice naturally engage a TH1 polarized immune response that resolves CL lesions, whereas BALB/c mice are prone to TH2 immunity, leading to disease progression (15). Furthermore, specific Leishmania strains can harbor Leishmania RNA virus 1, a potent stimulator of TLR3, resulting in induction of a type I interferon response (IFN-α/ß) and excessive inflammation promoting MCL lesion formation (16). Protective responses in VL involve TH1 responses to promote parasite clearance, while TH2 responses promote granuloma formation (17). Immune-deficient patients are particularly prone to life-threatening VL, where the spleen, liver, and bone marrow are infected (18). In addition to immune status and infecting Leishmania species and strain, leishmaniasis severity is also related to age or nutrition status (14). The role of immunity in leishmaniasis pathogenesis was recently reviewed in references (17, 19).

Trypanosoma cruzi

Trypanosoma cruzi is transmitted by triatomines, contaminated food or drink, through organ transplantation or blood transfusion, or transplacentally, and causes a life-long disease known as Chagas disease (CD) or American trypanosomiasis. The infection starts with an acute phase, where non-dividing trypomastigotes delivered in triatomine feces or in these other infection sources invade a broad spectrum of immune and non-immune cells including macrophage, dendritic cells, and cardiomyocytes. Trypomastigotes differentiate into amastigotes which replicate inside the host cells; prior to cell lysis the parasites differentiate back into trypomastigotes, which then disseminate to the extracellular microenvironment and blood (20). After control of parasite growth by the immune system, patients support persistent chronic infection with low parasite burden. Although most chronic CD patients remain asymptomatic, approximately 30% of infected individuals develop life-threatening chronic chagasic cardiomyopathy and, at lower frequency, gastrointestinal diseases (21). Control of infection in the acute stage relies on activation of both innate and adaptive effector immune responses. Recognition and phagocytosis of parasites by innate immune cells including macrophages, dendritic cells, and neutrophils result in canonical TH1 inflammatory response driven by TNF-α, nitric oxide, IL-12, and IFN-γ produced by T cells and natural killer cells (22). Additionally, TH17 is protective during acute T. cruzi infection through induction of NADPH oxidase and IL-21-dependent activation of CD8 T cells, resulting in reduction of parasite load and improvement of survival (23, 24). Antibody-mediated immunity plays a protective role in control of in vivo T. cruzi infection; however, antibody production does not completely eliminate the infection (25, 26). Also, B lymphocytes may perform regulatory functions on CD4⁺ T cells during T. cruzi infection. Trypomastigote pre-stimulated B cells induce apoptosis and reduce cytokine production in CD4⁺ T cells (27). For more detailed information on the immune response to T. cruzi, see reference (28). Thus, patients with immune deficiency, such as resulting from AIDS, organ transplants, or chemotherapy, present with reactivation Chagas disease, with elevated parasitemia, cerebral lesions, or meningoecephalitis, and myocarditis, with high mortality (22, 29).

Salivarian trypanosomes

Trypanosoma brucei gambiense and Trypansosoma brucei rhodesiense are the causative agents of human African trypanosomiasis (HAT). Trypanosoma brucei brucei is commonly used as a model system for HAT but is also a causative agent of animal trypanosomiasis in livestock and wild animals, along with Trypanosoma congolense, Trypanosoma evansi, Trypanosoma vivax, and other, less-studied, Trypanosoma species. Unlike HAT, animal trypanosomiasis has a much broader geographic distribution. Symptoms of HAT are non-specific when the parasites first enter the hemolymphatic system (stage 1): fever, fatigue, headaches, etc. Once parasites enter the central nervous system (meningoencephalitic stage), severe symptoms arise, including sleep and circadian rhythm disruption, seizures, coma, and ultimately death. Animal trypanosomiasis symptoms are predominantly wasting and anemia, with neurological symptoms much rarer (30). Unlike all the other agents described in this review, these parasites are exclusively extracellular. T. b. brucei and T. b. gambiense have tropism to the circulatory system and lymphatics (stage 1 disease stage) and the brain (meningoencephalitic stage), as well as the skin, adipose tissue, heart, and testes. Additional reports also provide evidence for colonization of multiple other tissues. In contrast, T. congolense primarily remains within the vasculature, but causing damage to most major organs. Like T. congolense, T. vivax and T. evansi cause pathological lesions in most major organs but with reports of extravascular tropism (31, 32). A robust B cell response is required to restrict T. brucei infection (33) though this is counterbalanced by parasite antigenic switching and parasite-induced T helper cell and B cell dysfunction (34). Readers are referred to a recent review for more details on the role of immune responses in these diseases (35).

APICOMPLEXAN PARASITES

Apicomplexan parasites are a diverse phylum of obligate intracellular parasites named for conserved apical organelle and cytoskeletal structures that facilitate cell entry into the host cell. The apicomplexans that are most pervasive in humans and best studied are Plasmodium sp., the causative agent of malaria, T. gondii, which is notable for its diverse host range and ability to infect most nucleated cells, and Cryptosporidium sp., the second leading cause of diarrhea in humans (36, 37). These parasites undergo dramatic structural transformations and changes in gene expression depending on metabolic and inflammatory environmental cues from the host. These parasites can directly stimulate toll-like receptors, leading to a canonical NFkB-driven type 1 immunity characterized by TNF, IL-6, nitric oxide, and IL-12 release from myeloid cells. IFNy, released from T cells, is critical to restrict acute infection and limit parasite load in the blood (parasitemia) (38, 39). While a robust antibody response is generated toward these parasites (40), the protective functions of this response are less well understood, as detailed below.

Plasmodium

Similar to kinetoplastids, Plasmodium species are transmitted via an insect vector, in this case Anopheles mosquitoes, which support the sexual recombination of Plasmodium gametes and the transmission of sporozoites. Sporozoites are transmitted when mosquitos take a blood meal. They rapidly infect liver cells and develop into schizonts. Schizonts rupture, releasing merozoites. Malaria results from subsequent merozoite infection of erythrocytes, leading to inflammation, anemia, and, if poorly contained, central nervous system infection that can be lethal. The virulence strategies and effector biology of Plasmodium are shaped by interaction with the immune system of the mammalian host as well as the mosquito, as reviewed excellently elsewhere (41, 42). Immunosuppression, particularly during HIV/AIDS, can increase prevalence of anemia and transplacental transfer of the parasite (43). There are at least 156 named species of Plasmodium. P. falciparum, P. ovale, P. vivax, and P. malariae are human-tropic species. However, the macaque parasite P. knowlesi can zoonotically infect humans. Due to this relatively stringent species tropism, animal models of malaria use P. berghei, P. yoelii, or P. chabaudi to understand selective aspects of human disease (44). P. berghei is often used as an acute, lethal model of tissue pathology (liver, spleen, lung) and cerebral malaria (44–47). P. yoelii has been used to model lethal infection at high doses or resolving infection at lower inoculum. P. chabaudi is notable as a model for cyclical infection characterized by periods of low parasitemia followed by cycles of parasite recrudescence and high parasitemia similar to what is often observed in human disease (48). The liver, which supports asexual Plasmodium replication for approximately 1 week, may be a bottleneck to systemic infection that can be controlled by CD8 T cell responses. The antibody response plays an important role regulating the cyclical nature of malaria by blocking parasite invasion of RBCs, inducing antibody-dependent killing of infected cells, and through opsonization and clearance of infected RBCs. However, these processes can be thwarted by surface antigen switching (49–51). While the TH1-response promotes parasite restriction, it also contributes to cytotoxicity, evident in the important role IL-10 plays in limiting immunopathology (38).

Toxoplasma gondii

Toxoplasma gondii has similar cell biology, invasion mechanisms, and effector biology as Plasmodium. However, T. gondii’s host range and transmission strategy are distinct. Rather than capitalizing on an insect vector, T. gondii’s definitive hosts are feline species. Cats shed environmentally stable oocysts which differentiate into sporozoites when consumed by an incredibly broad range of euthermic “intermediate” hosts, including avian species, livestock, rodents, and humans. The sporozoites invade the small intestine, possibly via epithelial cells, and transition into tachyzoites. Tachyzoites are capable of invading most nucleated cell types by injecting invasion machinery into the target cell. This is mediated by hundreds of effectors secreted from specialized parasite organelles into the host cell which facilitate entry, nutrient acquisition, and immune evasion (39). This remarkable pathogenesis strategy allows T. gondii to grow in stromal cell types and migratory immune cells which traffic the parasite to tissue cell types that support cyst formation and chronic infection, namely, the brain, cardiac, and skeletal muscle (52). Primary infection during pregnancy puts the developing fetus at risk due to its limited immune system. Stress and immune suppression can trigger T. gondii recrudescence, which can be lethal in the settings of AIDS, organ transplant, or some forms of chemotherapy (36, 52). In immunocompetent individuals, the acute phase of infection may lead to self-resolving, flu-like symptoms. However, the parasite can persist in neurons and muscle cells for the life of the host, gradually decreasing in numbers if the host immune system is intact (53). In this way, the distinct acute and chronic phases of infection are associated with distinct metabolic states of the parasite and unique metabolic constraints of the tissues that are infected at each phase (54). The metabolic state of the host is also linked to immune demand. A canonical TH1 immune response is necessary for parasite control, with IFNy playing important roles in cell autonomous immunity and the CD8 T cell response (55). Notably, parasite-specific IgG remains elevated in the serum throughout chronic infection, indicative of sustained priming of the immune response by the parasite (56). The limited metabolic studies on human patients look at chronically infected individuals with sustained T. gondii­-specific antibody responses (57–59).

Cryptosporidium

Cryptosporidium was first identified in murine gastric mucosa by Tyzzer in 1907 (60). Most Cryptosporidium species are host-adapted. Although 20 of the 26 named Cryptosporidium species have been isolated from humans, Cryptosporidium hominis and C. parvum zoonotically transferred from cows are the major contributors to human disease (61, 62). The global prevalence of Cryptosporidium infection is estimated at 7.6% of the general population (37, 63), with a higher incidence rate in resource limited countries (37). Unlike T. gondii and Plasmodium, Cryptosporidium sp. undergo gametogenesis and sexual recombination in a single host. Dissemination occurs when oocysts are shed into the water for oral transmission to the next host. As few as 10 oocysts are thought to be infectious (63). Cryptosporidium secretes effectors that are evolutionarily conserved with T. gondii that facilitate vacuole formation in the luminal face of small intestinal epithelial cells (64, 65). In immune-competent hosts, Cryptosporidium infection is primarily limited to the intestine, resulting in self-limiting diarrheal disease (66). Similar to T. gondii infection, commensal microbiota and parasite sensing play a role in eliciting a type I immune response which is necessary for host survival (67). In severe cases, particularly among immune-compromised patients, the parasite can translocate to the biliary tract, liver, lung, or pancreas, causing malabsorption, wasting, and life-threatening infection (66). There is evidence that even self-limiting infections in the GI tract can skew host metabolism (68). However, the molecular and biochemical mechanisms remain unclear. Similar to other apicomplexans, TH1 immunity is critical, but IL-18 produced by intestinal epithelial cells appears to play a critical role priming immunity in this niche (69). In both Cryptosporidium and T. gondii infection, there is robust evidence supporting a model where immune priming by commensal microbiota plays a dominant role initiating the immune response even though parasite components can directly ligate TLRs (70, 71).

WHY COMPARE HOST METABOLIC RESPONSES TO APICOMPLEXAN AND KINETOPLASTIDS?

Apicomplexan parasites and kinetoplastids diverged approximately 100 million years ago (72). Despite genome streamlining, many core metabolic demands remain conserved across these species. For example, both classes of pathogen are auxotrophs for a number of essential amino acids, which much be scavenged from the host (73–77). In both cases, during acute infection and recrudescence, there is tremendous demand for host lipids to fuel membrane biogenesis (78–81).

Over evolutionary time, host adaptation has driven the expansion of genus and species-specific effector gene families and pathogenesis strategies that reflect unique solutions to acquire metabolic substrates from the host. As discussed above, trypanosomes, Leishmania and Plasmodium transmission is transdermal, mediated by insect vectors (20, 30, 41, 82). However, T. cruzi is unique compared to these other pathogens in that it is deposited in the vector feces rather than the vector mouthparts. This has unique bearing on access to target tissues and interactions with immune cells during the onset of infection and recrudescence (83). T. brucei has evolved nutrient scavenging and immune evasion strategies to remain extracellular throughout infection, exploiting muscle, adipose tissue, and reproductive organs as sites for persistent infection and antigen variation (84). Plasmodium rapidly enters the liver, then persists within red blood cells, which have unique metabolic constraints due to their absence of nuclei and role in oxygen transport (85). Unlike T. brucei, Leishmania, and Plasmodium, another major mode of transmission for T. cruzi is orally, through contaminated food or drink, a similarity to T. gondii and Cryptosporidium (86). Despite their evolutionary distance, T. cruzi and T. gondii have similar pathogenesis strategies, as both organisms infect nucleated leukocytes in acute infection and can cause chronic infection in cardiac muscle (20, 36). However, T. gondii enters circulation through the small intestine and has a broader host cell tropism at acute and chronic phases of infection. Moreover, colonization of muscle and neurons is necessary for transmission (3). Cryptosporidium effectors are often conserved with T. gondii and Plasmodium, but Cryptosporidium, unlike T. gondii, completes its entire life cycle in the small intestine of the same host (63). Over time each of these parasites has evolved to exploit nodes of immune and metabolic signaling to meet their infection strategy and metabolic needs so that comparing these mechanisms can teach us about how these host pathways are regulated.

A challenge in understanding protozoan biology has been limited access to techniques for genetic manipulation and informatics tools to assemble genomes for parasites that use tandem repeat genes as part of their pathogenesis strategies. T. gondii was the first of the protozoan parasites discussed here to have robust forward and reverse genetic tools (52). T. gondii’s similarity to Plasmodium sp. has led to intense focus on the metabolic demands of the parasite as a lens to view changes in host metabolites that limit or facilitate intracellular growth. In contrast, genetic manipulation has proven more challenging in T. cruzi but is now achievable [e.g., reference (87)]. As we are entering an era where parasite genetic and metabolomic tools are becoming widespread, this is an interesting time to revisit what has been learned from these parasitic infection models in a way that can be cross-informative.

METABOLIC PATHWAYS ALTERED DURING KINETOPLASTID AND APICOMPLEXAN INFECTION

Amino acid metabolism

Amino acid restriction, nutritional immunity, and immune response polarization

Amino acids are critical precursors for host and parasite polyamine synthesis. Apicomplexans and kinetoplastids are auxotrophic for several amino acids (88), indicating that restricting amino acid availability may be an important means of nutritional immunity. Plasma amino acid depletion is a pervasive symptom of acute Plasmodium infection in vertebrates (89, 90) that plays a causal role in severe malnutrition (91). Systemic inflammatory signals can induce hepatocytes to take up circulating amino acids (92), followed by amino acid import by Plasmodium (93); the best studied of these are tryptophan, arginine, and glutamine (Fig. 1) (90, 94, 95). Depletion of several plasma amino acids has also been reported in T. b. brucei-infected mice up to 33 days post infection and in T. b. rhodesiense-infected humans (both disease stages combined) (96, 97), and tryptophan levels are reduced in the cerebrospinal fluid of T. b. gambiense stage 2 patients (98). Several amino acids, such as aspartate, methionine, lysine, and phenylalanine, were depleted in the hearts of Chagasic cardiomyopathy patients compared to healthy controls (99). This is a valuable study due to its analysis of these hard-to-access human samples. However, differences in metabolism due to sampling between those groups (at the time of cardiac explant for Chagas disease patients vs post-mortem for healthy controls) cannot be excluded. A further confounder is differences in patient age between the groups. The serum from active visceral leishmaniasis disease patients is deficient in methionine, homoserine, threonine, and histidine (100). Hamsters infected with L. donovani showed mostly a decrease in amino acids and peptides in the spleen at 12–14 weeks post infection (101). In this model, parasite burden in the spleen at this late-stage timepoint remains consistent with parasite burden 5 weeks post infection (102). These results concur with the upregulation of serine, cysteine, and tyrosine catabolic enzymes in the hamster spleen at 28 days post infection (103) though this study did not investigate later time points. Comparatively little is known in Cryptosporidium. However, increased serine and glutamine consumption was observed in C. parvum-infected HCT-8 cells, and pharmacological inhibition of glutaminolysis impaired parasite growth, demonstrating a causal association (104).

Fig 1.

Amino acids and immune response modulation. (A) Glutamine can fuel the TCA cycle and glycolysis. The latter plays an important role in priming TH1 immunity and B cell responses that clear protozoan and apicomplexan parasites. Glutamine depletion conditions promote regulatory T cell responses (Treg) and limit TH17 responses in the intestine and brain, which can promote tissue pathology. (B) Classical activation upregulates iNOS which converts arginine to nitric oxide (NO), the precursor for reactive nitrogen species (ROS), and ROS which can be directly parasiticidal and cytotoxic. In contrast, “alternatively activated” macrophages are more reliant on beta-oxidation and use arginase to convert arginine to ornithine and urea. Arginase activity can be promoted by cruzipain synthesis in T. cruzi infected phagocytes. Alternative activation is associated with TH2 immunity and wound healing environments regulated by innate and adaptive immune cell cross talk with stromal cells in the tissue that promote fibroblast activation and extracellular matrix development. (C) Indolamine can be metabolized into kynurenine, which is associated with pro-resolution responses and the inhibition of TH1 and TH17 immunity. (D) Tryptophan can also be metabolized into kynurenine by the host or by T. brucei into indolepyruvate to inhibit TH1 and TH17 responses. Tyrosine conversion to hydroxyphenylpyruvate blunts glycolytic metabolism and the “classical activation” of antigen-presenting cells including macrophage (Mo) and dendritic cells (DC). Figure created with BioRender.com.

For most studies, it is unclear if the reduction in amino acids reflects hyper-catabolism or reduced biosynthesis. In support of the first possibility, reduced phenylalanine incorporation into proteins in the liver, brain, kidney, and heart was observed in voles infected with T. b. gambiense at 6 weeks post infection, along with increased circulation and excretion of downstream metabolic products, though this older study may not have fully removed circulating phenylalanine during tissue studies (105). Glutamine, alanine, serine, valine, and histidine are glucogenic amino acids that can fuel glycolysis (as discussed below). Tryptophan, tyrosine, isoleucine, and threonine can be converted to glucose to drive glycolysis or acetyl-CoA. Acetyl-CoA and the lysine/leucine derivative acyl-CoA can drive mitochondrial ATP synthesis or production of ketone bodies, which may be in high demand due to sickness behaviors like anorexia (106). For example, acute T. cruzi infection increased leucine, isoleucine, valine, and their catabolic metabolites such as alpha-keto acids in the heart and plasma in mice (107). Increased serine consumption was observed in C. parvum-infected HCT-8 cells, which may be required as an essential building block for ceramides and phospholipids (108).

The pressures driving amino acid depletion may reflect host metabolic needs during infection, as well as parasite demand, especially during acute infection when parasite burden is high. Phenylalanine was increased in the plasma of humans infected with T. b. rhodesiense (stage 1 and stage 2 combined) (97). Notably, glutamine can fuel the TCA cycle of both host and Plasmodium parasites (109). Using isotope-labeled glucose and glutamine, glutamine was shown to be the preferred input of the TCA cycle in both asexual and sexual blood stage of P. berghei and the asexual blood stage of P. falciparum (109). Low plasma glutamine is associated with severe malaria in patients infected with P. falciparum and could represent depletion by the parasite (110). In contrast, caloric restriction and mobilization of amino acid substrates for host metabolism trigger nutrient sensing pathways in Plasmodium, associated with reduced parasite load and dissemination during blood stage (111). Less is known about the source of amino acids in liver stage Plasmodium infection; however, inhibiting cationic amino acid uptake impairs liver parasite burden in vivo (112).

While amino acid restriction may limit parasite growth, it can also impair the immune response. Treating patients with a glutamine analog that cannot be metabolized prolonged host survival and increased parasite-specific immune cells in mice with late-stage cerebral malaria (113). However, experimental depletion of glutamine in patients also exacerbates arginine deficiency in malaria (114), indicating that the overall benefit of modulating this pathway for host outcome and its causality in disease pathogenesis remains to be determined. Serine restriction reduces IFNγ production and CD4 and CD8 effector T cell expansion, ultimately impairing primary and memory T cell responses (115). Glutamine starvation shifts Th1 immunity toward a Treg environment (116) and impairs TH17 cell differentiation (Fig. 1A) (117). Glutamine is depleted in murine L. major cutaneous lesions (118) and in the serum of active visceral leishmaniasis patients (100), which may impair the productive TH1 response necessary to protect against leishmaniasis (Fig. 1A) (119). Consistent with the model, supplementing glutamine in combination with antiparasitic miltefosine improved CD8 cell recruitment to the spleen, enhancing IFNγ production and limiting IL10-producing cells, and improved L. donovani clearance (Fig. 1A) (120). These results confirm a causal role of glutamine depletion in leishmaniasis pathogenesis. The importance of glutamine metabolism may not be limited to a productive T cell response: oral glutamine supplementation during acute P. yoelii infection was sufficient to boost acute and memory B cell responses and promote host survival (121).

In addition to these direct immunomodulatory effects, amino acid-derived metabolites can regulate immune signaling. For example, the host tyrosine product hydroxyphenylpyruvate and the T. brucei tryptophan metabolite indolepyruvate can both inhibit expression of pro-inflammatory cytokines pro-IL1β, IL-6, and TNFα, prostaglandin production by macrophages, and dendritic cell maturation (122–125). When immune cells encounter pro-inflammatory signals, indolepyruvate inhibits the shift to glycolytic metabolism (125).

Arginine metabolism in nutritional immunity and synthesis of reactive nitrogen species

Arginine metabolism is a critical immunometabolic regulatory axis for parasite infection. Leishmania, T. cruzi, T. brucei, and apicomplexans are arginine auxotrophs and require the host amino acid for polyamine synthesis (73–77, 126). Depleting L-arginine reduces liver stage Plasmodium load and limits T. gondii growth in tissue culture, leading to cyst formation (75–77). L-arginine is also a substrate for host nitric oxide synthases (iNOS, eNOS, and nNOS) and arginase enzymes, which regulate antagonistic immune modulatory pathways (Fig. 1B). NOS enzymes metabolize arginine to citrulline and nitric oxide (NO), a reactive gas with diverse signaling functions and a precursor for reactive nitrogen species involved in parasite killing (Fig. 1B, green macrophages) (127–129). Arginase converts arginine to ornithine and urea. Arginase is transcribed by IL-4/STAT6 signaling, linking its activity to alternative macrophage activation, TH2/humoral immunity, and wound healing responses (Fig. 1B, blue macrophages) (130). Mouse dendritic cells infected with T. gondii are high in urea, ornithine, and downstream metabolites (131), and hypervirulent type I strains promote arginase activity via the effector protein Rop16, leading to an alternative macrophage phenotype that supports intracellular parasite growth. Notably, when Rop16 was expressed in a less virulent type II T. gondii strain, intestinal health improved after oral infection, indicating that the arginase-dependent pro-resolution phenotype can benefit the host as well as the parasite by limiting inflammation-induced toxicity (132).

This dual role for arginase extends to trypanosome infection. Treating macrophages with cruzipain, a T. cruzi virulence factor, increased arginase activity and polyamine synthesis, favoring intracellular T. cruzi growth (Fig. 1B) (133). In the heart and plasma of acutely infected mice, levels of arginine and citrulline (iNOS-derived) were depressed, while cardiac ornithine and urea (arginase products) were increased (134). Likewise, arginine was decreased in the hearts of Chagas disease patients with end-stage heart failure compared to healthy controls (99). Arginine supplementation not only improved heart function and host survival but also increased cardiac iNOS levels, so it is unclear if the benefit was conferred by the wound healing functions of arginase or the parasite killing downstream of iNOS (Fig. 1B) (134). Multiple cattle breeds infected with T. congolense have elevated expression of arginase in blood (135). However, plasma ornithine was lower in stage 2T. b. gambiense patients compared to stage 1 (98).

Citrulline, ornithine, and polyamines are also higher in the plasma of diffuse cutaneous leishmaniasis patients compared to localized cutaneous and mucocutaneous leishmaniasis (136). The presence of both iNOS and arginase products may indicate context-dependent elevation of both enzymes in diffuse cutaneous leishmaniasis. In mouse models of cutaneous leishmaniasis, the importance of arginase in disease pathogenesis depends on host genetics: knocking out arginase had no effect on disease progression in L. major infection of the resistant C57BL/6 model which is characterized by TH1 polarized immunity (Fig. 1B) (137). In contrast, arginase deficiency exacerbated disease progression in the susceptible, TH2-polarized BALB/c mouse (138). In human cutaneous leishmaniasis, arginase 1 transcript levels are depressed, consistent with a C57BL/6-like TH1 environment (139).

In addition to a role for the L-arginine/iNOS/NO axis in parasite clearance from infected cells, reactive nitrogen species are the most potent molecular regulators of vasodilation via eNOS activity in erythrocytes and endothelial cells, which can also impact the outcome of infection (90, 140, 141). Low plasma arginine has been associated with low birth weight and infant mortality in a cohort of pregnant women with malaria (142). In P. berghei-ANKA mouse models of malaria, dietary arginine supplementation rescued the development of placental blood vessels, fetal weight, and pup viability (142) and improved cerebral blood flow, reduced vascular leak, and host survival of cerebral malaria (143–145), indicating a causal role in disease pathogenesis. Despite this well-established relationship, the physiological mechanisms of arginine remain controversial. Rubach et al. found that children with cerebral falciparum malaria have similar arginine flux as healthy children by tracking the metabolism of isotope-labeled arginine. This suggested that hypoargininemia and low NO production in malaria patients are due to limited arginine synthetic precursors rather than hypermetabolism of arginine (146).

Kynurenine metabolism and aryl hydrocarbon receptor-mediated immune suppression

Kynurenine is produced from tryptophan by tryptophan 2,3-dioxygenase (TDO) and indoleamine 2,3-dioxygenase-1 (IDO) (Fig. 1C). Kynurenine and its derivatives are ligands for host aryl hydrocarbon receptor (ARH), a transcription factor that regulates cell homeostatic functions and suppresses immune responses. IDO expression is upregulated by IFN-γ and TNFɑ to limit proinflammatory transcriptional responses (147, 148). Kynurenine and other tryptophan metabolites block IFN-γ and IL-17 synthesis and enhance anti-inflammatory T regulatory cell differentiation, IL-10, and TGFβ production (Fig. 1C) (149).

Activation of the kynurenine pathway is conserved across apicomplexan and kinetoplastid infection. Low tryptophan and increased kynurenine and kynurenic acid are observed in the serum of active visceral leishmaniasis patients (100). IDO is induced in hamsters infected with L. donovani 28 days post infection (103) and in both human and mouse cutaneous leishmaniasis (139, 150). Likewise, kynurenine is increased in the heart of T. cruzi-infected mice during acute infection (107). Using a chronic model of P. chabaudi infection, Lissner and colleagues found that peak parasitemia (5–12 days post infection) corresponds with a severe reduction of serum tryptophan followed by a spike in the levels of the downstream AHR ligands kynurenine, kynurenate, and quinolinic acid (Fig. 1C and D) (89, 140). Given the anti-inflammatory effects of these metabolites (147–149), elevated kynurenine and IDO may contribute to parasite persistence.

Inducing this metabolic program may alternatively protect the host from pathological inflammation. Ahr^−/− mice were susceptible to acute P. chabaudi infection; however, this was not due to parasite overgrowth. Instead, animals presented with elevated bilirubin, a marker of kidney injury conserved in human malaria. Tryptophan metabolites regulate the severity of acute kidney failure and neurological pathology during malaria (89, 90). While there have been reports associating the kynurenine/ADH axis with cerebral malaria, it is not clear if tryptophan metabolites are causing pathology or are a signature reflecting a need for immune-inhibition in this sensitive tissue (151, 152). Hyperbaric oxygen treatment of P. berghei-infected C57BL/6 mice improved cerebral hypoxia and reduced IDO and kynurenine levels (153). In malaria patients, the cerebrospinal fluid level of quinolinic acid (a downstream product of kynurenine metabolism) is positively correlated with the severity of neurological pathologies (154). Another counter-intuitive observation is the negative correlation between the abundance of TDO transcripts and parasite burden in human cutaneous leishmaniasis lesions (139). Leishmania and Apicomplexans are tryptophan auxotrophs, and there is compelling evidence to suggest that although kynurenines inhibit inflammatory responses, this may be balanced by nutritional immunity imparted by starving intracellular parasites of tryptophan (155).

Nucleotide, nucleoside, and nucleobase metabolism

Nucleosides are purine nucleobases (adenine and guanine) or pyrimidine nucleobases (cytosine, thymine, and uracil), covalently bound to a pentose. Addition of at least one phosphate group generates nucleotides (Fig. 2). Overall, these metabolites have distinct functions inside the cell, where they store information (e.g., DNA and RNA bases), transfer energy (e.g., ATP, ADP), act as second messengers (e.g., cyclic(c)AMP, cGMP), and can activate cytosolic immune sensors (e.g., cyclic-di-GMP); vs outside the cell, where they function as ligands for cell surface receptors (156, 157). Their extracellular functions make them important modulators of metabolism and immune function. In particular, ATP ligation of P2 × 7R initiates pyroptotic cell death (156). ATP signaling can also promote T cell activation and immune cell chemotaxis, while extracellular adenosine signaling is predominantly anti-inflammatory (158). Ectonucleotidases CD39 and CD73 jointly degrade extracellular ATP to produce adenosine. Ectonucleotidase inhibition, thus, led to higher IFN-γ, TNF-α, and IL-12 production, lower IL-4, IL-10, and TGF-β and improved parasite clearance at 4 weeks post infection from liver and spleen in mouse models of L. donovani infection (Fig. 2A) (159).

Fig 2.

Nucleoside metabolism in parasite infection. Purines (tan) and pyrimidines (purple) serve as intracellular metabolites and extracellular signaling molecules. Kinetoplastids (A, B, and E) and apicomplexans (C and D) are purine auxotrophs that must acquire these metabolites from the host. Nucleoside balance is frequently dysregulated during infection depending on tissue type and time during infection, which may be related to inflammatory response and/or parasite burden. A2AR, adenosine A2A receptor. Figure created with BioRender.com.

Altered purine and pyrimidine bioavailability, frequently reduced, is a common feature of infection. This could be directly due to parasite scavenging since kinetoplastids and apicomplexans are purine auxotrophs (160). Xanthine dehydrogenase, which catalyzes purine catabolism, was upregulated in the spleens of hamsters infected with L. donovani at 28 days post infection (Fig. 2A) (103). Uridine and uracil are also low in the serum of active visceral leishmaniasis patients (100). Purine levels are lower in the skin than in the liver (161), which may require different parasite adaptations (162) and likewise differentially shape immune responses. Adenine levels are depressed in the mouse heart tissue following acute T. cruzi infection (107), and the chronically infected heart is depleted in adenosine, inosine, and hypoxanthine, accompanied by increased levels of urate, a breakdown product of purine (Fig. 2B) (163). However, factors beyond direct parasite scavenging are involved since this purine depletion persists even following antiparasitic treatment (163).

Acute spikes in nucleotide metabolism may represent host demand for immune signaling or compensation for scavenging by the parasite. For example, dual metabolic profiling of fibroblasts infected with T. gondii showed that nucleotide abundance and levels of the purine and pyrimidine synthase enzymes methylenetetrahydrofolate synthase and phosphoribosylpyrophosphate synthetase were increased from 6 to 48 hours post infection (54). In vivo, purine levels are elevated in the brain 6 days post infection with the hypervirulent Type I strain and Type II T. gondii (Fig. 2C). However, by 30 or 60 days post infection, the levels of ATP, ADP, adenosine, inosine, and xanthine had decreased in Type II infection, suggesting availability is mobilized during the acute inflammatory response and/or parasite demand, which is subsequently decreased in chronic infection (164). Notably, inosine has emerged as a negative regulator in inflammatory responses via signaling through the adenosine A2A receptor (Fig. 2); however, a direct immune down modulatory mechanism has not been explored in T. gondii infection (165). Transcripts regulating nucleotide metabolism were upregulated in brains infected with T. gondii at 10 and 38 days post infection (166). A spike in pyrimidine levels has also been reported in P. chabaudi infection, concomitant with peak parasite load at 5–10 days post infection, which resolved by 20 days post infection (Fig. 2D) (140). RNA-seq analysis of adipose tissue from T. brucei-infected mice showed increased purine and pyrimidine biosynthesis at day 6 post infection compared to day 0 (Fig. 2E) (167), with the caveat that RNA analyses do not always reflect protein levels and cannot inform on enzyme activity.

Beyond immunomodulatory effects, altered nucleotide and nucleoside levels may also signal to promote anti-parasitic metabolic shifts. For example, higher levels of AMP and adenosine are observed at the heart apex in uninfected mouse hearts (Fig. 2B) (168). AMP is an activator of the enzyme AMP-activated protein kinase (AMPK), which, in turn, inhibits T. cruzi growth (169). Thus, pre-existing elevated AMP levels at the heart apex may restrict T. cruzi growth, leading to the lower parasite burden observed at this site (168), though this awaits determination of causality, for example, by treatment with AMPK agonists and antagonists, followed by assessment of parasite spatial distribution.

Lipid metabolism

Lipid metabolism and storage play a critical role in nutrient acquisition by parasites and drive immune cell functions during disease. As discussed below, IL-4 and STAT6, canonical TH2 immune signals, promote lipid oxidation by mobilizing peroxisome proliferator activated receptors (170, 171). Lipid stores are a critical source of energy drawn upon during anorexia, a classical sickness behavior associated with high inflammatory mediators released during peak parasitemia, including IL-1 (formerly named anorexin) (172). There is increasing evidence that many parasites can benefit from short- or long-term residency in adipose depots, which may be related to the alternative-activation-type inflammatory environment and the availability of metabolic intermediates at these sites (Fig. 2A, bottom) (173).

Indeed, adipose tissue is a reservoir for T. cruzi during acute and chronic infection (174). T. cruzi infection alters adipocyte structure and physiology in vivo. T. cruzi-infected hearts displayed accumulation of large lipid bodies during acute infection in mice. Interestingly, chemical ablation of adipocytes in acute CD in mice resulted in the accumulation of small lipid bodies, increased parasite load, immune cell infiltration, and cell damage in cardiac tissue, associated with decreased survival (Fig. 3A, top). On the other hand, chronic T. cruzi infection caused an imbalance between adipogenesis and lipolysis which further increased by fat ablation. Furthermore, loss of fat cells led to increased cardiac damage during chronic CD (175). These results indicate that changes in adipose tissue metabolism control CD pathology. Adipose tissue also serves as a major site of infection for T. brucei (176). As visceralizing Leishmania species directly infect the liver, they directly cause major alterations to lipid metabolism, including depleted serum cholesterol, elevated serum triglycerides, elevated spleen ceramides and acylcarnitines, and mostly elevated liver acylcarnitines 12–14 weeks post infection (101, 177, 178). T. gondii has also been observed replicating in epigonadal and mesenteric adipose depots, and acute infection is associated with anorexia and adipose depot wasting (172, 179). Major changes in lipid profile are, thus, commonly observed across these parasite infections though altered breakdown vs altered biosynthesis is often hard to differentiate.

Fig 3.

Dysregulation of lipid metabolism during kinetoplastid and apicomplexan parasites infections. Parasitic infections mainly affect host lipolysis/beta oxidation and lipid biosynthesis/lipid scavenging. (A) T. cruzi infection as an example of lipolysis/beta oxidation perturbation. T. cruzi trypomastigotes infect a range of tissues including adipose tissue and cardiomyocytes. Top, infection is associated with elevation of long-chain fatty acids and acylcarnitines in the heart during acute infection, suggesting beta oxidation modulation. Bottom, amastigote replication in host cells leads to infiltration of immune cells and production of inflammatory cytokines including TNFα in adipose tissue which results in increased lipolysis. AAM, alternatively activated macrophages. (B) Parasitic scavenging of host lipids plays an important role in the perturbation of host lipid metabolism. Left, T. gondii and P. falciparum rely on host cholesterol while they can synthesize fatty acids and phospholipids. Intracellular T. gondii uptakes host cholesterol endocytosed by LDL receptor into parasitophorous vacuole. P. falciparum also imports host cholesterol by Niemann–Pick Type C1-Related 1 located in the parasite membrane. Middle, upregulation of cox2 leads to increases in prostaglandin levels in T. gondii, and Leishmania infection, while Plasmodium and T. brucei can also directly produce prostaglandins. Right, although kinetoplastid parasites are able to synthesize their own ergosterol, they uptake other host lipids including triglycerides, diglycerides, and some glycerophosphocholines. Fatty acids are metabolized through the acylcarnitine pathway. NEFA, non-esterified fatty acids. TGA, triglycerides. LCFA, long-chain fatty acids. Figure created with BioRender.com.

Lipid oxidation, parasite proliferation, and tissue pathology

Lipid oxidation is a major catabolic pathway. A key aspect of lipid catabolism is the mobilization of lipids from tissue stores, especially from adipose tissue. In T. b. brucei-infected mice, lipolysis is driven by infiltrating lymphocytes and helps control adipose tissue parasite burden, likely through cytotoxic effects of released free fatty acids, and thus preventing host mortality (Fig. 3A) (180). However, the level of circulating lipids is also lowered in mice, suggesting that lipid oxidation is likely high (Fig. 3A, bottom) (96). In contrast, rabbits infected with T. b. brucei have impaired lipid uptake from circulation at 3–4 weeks post infection though this study only used three animals/group (181). Acute T. cruzi infection in mice causes lipolysis accompanied by a reduction in the adipokines signaling adipogenesis and elevation of TNF-ɑ and IFN-γ in white adipose tissue (Fig. 3A) (182–184). Mice lose 60%–80% of their fat mass during acute CD, which can be rescued by administration of TNF-ɑ (but not IFN-γ) neutralizing antibodies, demonstrating a causal link between immune responses and these metabolic changes (Fig. 3A) (185).

Metabolomic analysis has shown that long-chain fatty acids (LCFAs) are decreased in the plasma but increased by infection in cardiac tissue during acute CD in mice (Fig. 3A) (107). Also, acylcarnitines, which are necessary to shuttle cytosolic LCFAs into the mitochondria for lipid oxidation, are increased in heart tissue during acute CD (Fig. 3A and B) (107). These findings suggest perturbation of cardiac beta oxidation during acute CD and may be associated with the mitochondrial impairment observed in this disease in mouse models and in humans (186–190). Long-chain fatty acid oxidation is positively associated with T. cruzi amastigote growth in primary human fibroblasts (169), indicating that decreased flux through this pathway may protect against T. cruzi in vivo. Long-chain acylcarnitines are also associated with non-fatal outcomes in acute CD (168). Metabolic restoration following acute-stage carnitine treatment, in association with improved survival, causally links these findings to disease pathogenesis, with the caveat that metabolic changes beyond acylcarnitines were also observed subsequent to treatment (191). In contrast, several cardiac and serum acylcarnitine species and the expression of liver carnitine palmitoyltransferase-1 were depressed in mice chronically infected with T. cruzi (192, 193) though this appears to be dependent on the combination of parasite strain, mouse strain, and timepoint (163, 194). Evaluation of the effects of carnitine treatment on metabolism during chronic T. cruzi infection is ongoing. Likewise, chronic T. gondii infection in mice decreased serum acylcarnitines (195) and several acylcarnitines were decreased in the serum of active visceral leishmaniasis patients (100). In contrast, higher plasma acylcarnitines were observed in C57BL/6 mice acutely infected with P. chabaudi (140). This suggests differential impacts of infection on lipid oxidation depending on the infecting parasite and infection timepoint, with the specific pattern related to the presence or absence of cachexic status (195).

Lipid biosynthesis, host cell demand, and parasite scavenging: cholesterol

The importance of lipid acquisition for apicomplexan and kinetoplastid biology is highlighted by how many nodes of host lipid transport, metabolism, synthesis, and scavenging these parasites have evolved to manipulate (Fig. 3B) (78–81). A strict requirement for host cholesterol scavenging, which is necessary for plasma membrane integrity, may be the best-understood lipid dependency of apicomplexans. Some host cholesterol is incorporated into the T. gondii vacuole membrane during parasite entry (196). T. gondii subverts receptor-mediated endocytosis of LDL to acquire cholesterol (197) and P. falciparum expresses Niemann–Pick Type C1-Related 1 to import scavenged cholesterol throughout infection (Fig. 3B) (198). Humans chronically infected with T. gondii have higher total cholesterol and LDL than seronegative individuals (199). While it is tempting to speculate that this is driven by parasite demand, this is difficult to test. In mice, LDL was increased throughout infection and HDL was elevated at 5–6 weeks post infection (199). T. gondii is a cholesterol auxotroph that imports host lipids and competes with mitochondria for lipid metabolic substrates (200, 201). Thus, pathways that promote lipid synthesis and turnover may benefit parasite metabolism. Malaria patients have lower plasma levels of total cholesterol, HDL, and LDL than patients with other febrile diseases (Fig. 3B) (202, 203). Similarly, low cholesterol was also observed in a malaria cohort from Gabon as compared to other febrile diseases (204). Notably, liver stage Plasmodium requires its own lipid synthesis machinery even though lipid transcripts for lipid importers are expressed (205). Whether these shifts in host systemic metabolism are due to host and/or parasite demand or a mechanism of nutritional immunity remain to be explored.

Unlike apicomplexans, kinetoplastids synthesize ergosterol as their major membrane sterol although T. brucei bloodstream forms prefer to scavenge host cholesterol (Fig. 3B) (206). Nevertheless, cholesterol availability is also altered in kinetoplastid infection. Blood cells from cattle infected with T. congolense express high levels of cholesterol biosynthesis machinery (135) and cholesterol levels were elevated in cardiac tissue of chronic CD patients (184). In contrast, in human cutaneous leishmaniasis lesions caused by L. braziliensis (207), and in T. brucei infected adipose tissue (167), the expression of genes regulating cholesterol biosynthesis is depressed (Fig. 3B). These contrasting observations may reflect site-specific regulation of lipid availability or, in infected tissues, of cholesterol uptake and metabolism by the parasite.

Lipid biosynthesis, host cell demand, and parasite scavenging: other lipids

Low levels of circulating lipids and shifts in the relative abundance of lipid species may reflect decreased fatty acid biosynthesis by the host or the selective import of lipid species by the parasite. Plasmodium and T. gondii are capable of synthesizing fatty acids and phospholipids (78). However, these parasites also scavenge host fatty acids to fuel the massive demand for phospholipids during haploid division and to synthesize secretory vesicles containing effectors. Similarly, T. cruzi can synthesize its own lipids, but heavily scavenges host triglycerides, diglycerides, and some host glycerophosphocholines as sources for specific parasite lipid classes (208). Similar scavenging is performed by Leishmania (209, 210) and T. brucei (211). While this latter study is old, its findings remain valid.

Increased membrane biosynthesis due to parasite proliferation, as well as host repairs of invasion-induced damage, may be the cause of the elevated glycerophosphocholine levels commonly observed across kinetoplastid and apicomplexan infections (107, 118, 140, 163, 193–195, 212). Likewise, the elevated fatty acid biosynthesis observed in the spleen of hamsters acutely infected with L. donovani may be facilitating membrane biosynthesis (103). However, in CL lesions, fatty acid biosynthesis genes were decreased (207). In T. b. rhodesiense human infection (stage 1 and stage 2 combined), increased levels of several glycerophosphocholines and decreased levels of other glycerophosphocholines, lysophosphocholines, and sphingomyelin have been observed in plasma, contrasting with the other kinetoplastids (Fig. 3B) (97). Low phosphatidylcholine was also observed in plasma of T. b. brucei-infected mice compared to pre-infection controls at 7 days post infection (96), and fatty acid biosynthesis genes are depressed in T. brucei-infected adipose tissue (167).

T. gondii uses secreted effectors to upregulate host triacylglycerol synthesis enzymes and fatty acid binding proteins, leading to lipid droplet accumulation in host cells (Fig. 3B) (213). This may also be responsible for the observed depressed non-esterified fatty acids and triglycerides in the serum of chronically T. gondii-infected mice compared to uninfected animals (195). T. gondii can synthesize ceramides and phosphoethanolamines using three enzymes (TgSpt1, -2, and des) that require import of host dihydroceramide as a substrate (214). When these enzymes are knocked out, parasites can still import and utilize host sphingomyelin. To acquire lipids, parasites can subvert vesicle trafficking from the endo-lysosomal pathways and trans-golgi network into the intravacuolar network (214, 215). There is also evidence that the T. gondii may promote lipophagy to acquiring fatty acids, oleic and linoleic acid from the host (214, 215). In return, host cells compete for fatty acids by relocalizing mitochondria to the vacuole where they limit lipid uptake by the parasite and parasite growth (201). In acute and chronic murine T. gondii infection, the liver has increased sphingolipids and fatty acids (216). C57BL/6 mice with P. chabaudi infection have increased sphingolipids in the plasma during acute infection compared to uninfected mice, which gradually return to normal from 12 to 25 days post infection as parasite load declines (Fig. 3B) (140). Malaria patients have higher plasma triglycerides compared to healthy donors (202, 203). Similarly, high triglycerides were also observed in a malaria cohort from Gabon as compared to other febrile diseases (204). Systemic inflammation-induced hepatic tissue damage (217) may aggravate the dysregulated lipid metabolism in malaria because the liver critically regulates the lipid homeostasis in humans and mice. Indeed, systemic inhibition of pro-inflammatory TNF-α signaling significantly reduces liver triglycerides in P. yoelii-infected mice, suggesting a causal association (218). These results raise a hypothesis that malaria-induced pathologies in the liver may play a major causal role in the metabolic disorders in the host. Host lipid acquisition is necessary to support parasite membrane biogenesis during acute stages of T. gondii growth (200). Thus, manipulating systemic lipid metabolism in vivo may be a strategy for pathogenesis although this has not been tested directly. A similar pattern of inflammation and liver damage may also be responsible for the observed increase in serum triglycerides during L. donovani infection since these lipid levels were proportional to disease severity (Fig. 3B). However, their levels were not significantly correlated to circulating IL-1β, IL-6, or TNFα, with the caveat that serum cytokine levels may not fully reflect local tissue inflammation (178, 219, 220).

Of note, bulk tissue or whole-cell lipidomic analysis may only reveal lipid scavenging by the parasite if the extracted lipids are subsequently metabolized by the parasites. Likewise, loss of host lipids may be masked by parasite lipid biosynthesis if the same lipid structures are synthesized by both. Instead, fractionation of intracellular parasites vs host cells or subcellular metabolomics techniques, still in their infancy, are needed (221). Mass spectrometry imaging can also demonstrate the localized impact of infection on lipid metabolism, as recently implemented with L. mexicana and L. donovani (222, 223).

Lipid metabolites as signaling modulators of the immune response to parasites

In addition to storing energy, adipose tissues are endocrine organs secreting adipokines and inflammatory mediators that regulate the metabolism of adipocytes and other cells (224). Circulating lipids, especially non-esterified fatty acids, are immunomodulatory molecules with structure-dependent signaling functions (225). Oxidized lipid species, generated in the presence of reactive oxygen species, are enriched during infection and can positively and negatively regulate TLR signaling (226). Ceramide, increased during active T. b. rhodesiense infection (97), can be pro-inflammatory (227). Fatty amides were also increased in both stages of T. b. gambiense infection in biofluids (98).

Eicosanoids, a broad class of lipid mediators that includes prostaglandins (PG) and leukotrienes, are synthesized from diacylglycerol or membrane phospholipids and regulate immune response kinetics and polarization (228). T. gondii infection induces expression of cyclooxygenase 2 (COX2), which produces the precursor prostaglandin (PG) PGH₂, and PGE₂ in infected macrophages (229). Unlike bacteria and viruses, eukaryotic pathogens can themselves synthesize eicosanoids, which complicates our understanding of their roles in parasite infection. P. falciparum and T. brucei produce PGD₂, PGE₂, and PGF_2α (230, 231). PGD₂, which promotes sleep, is increased in the cerebrospinal fluid of stage 2 patients infected T. b. gambiense, and it is not clear if this is host or parasite derived (232, 233). Host PGE₂ is elevated in malnourished mice (zinc and iron-deficient), contributing to high CCR7 levels that promote the trafficking of L. donovani-infected monocytes to the spleen and suggesting that PGE₂ may contribute to lesional spread (234). Diffuse cutaneous leishmaniasis is associated with elevated plasma PGE₂ compared to localized cutaneous leishmaniasis, visceral leishmaniasis, and uninfected patients, suggesting that prostaglandin regulation may be context- or tissue-specific (235, 236). Consistent with this model, COX2 is elevated within the skin lesion of patients with diffuse cutaneous lesions compared to localized cutaneous leishmaniasis manifestations though this study relied on non-quantitative immunohistochemistry studies (236). COX2 is also elevated in the spleens of L. donovani-infected mice (237). Persistent elevation in plasma eicosanoids is associated with treatment failure in cutaneous leishmaniasis, and pre-treatment eicosanoid profile is predictive of treatment outcome (238). In contrast, PGE₂ and COX2 are elevated in the serum of healthy, malaria-exposed children but reduced in children with acute malaria (239). Following P. yoelii infection, dendritic cells migrate to the spleen and secrete TGF-β, PGE₂, and IL-10. Inhibition of PGE₂ and TGF-β is sufficient to boost CD8 T cell immunity to sporozoites and protect the host from re-infection (240).

Glycolysis and the pentose phosphate pathway

Kinetoplastids

Glycolysis is the oxidation of glucose to pyruvate, which is further metabolized to acetyl-CoA and finally released as CO₂ through the TCA cycle (Fig. 4 and see below). Pyruvate can alternatively be fermented to lactate to regenerate reducing equivalents (NAD⁺). In this way, lactate production may reflect conditions of high energy demand, and/or low oxygen availability. Mice infected with T. b. brucei had elevated urine and serum lactate as well as serum glucose, possibly reflecting a differential balance between glucose catabolism (Fig. 4A) (generating lactate) and anabolic reactions (synthesizing glucose) between tissues (96). Likewise, whole blood cell RNA-seq analysis showed that glycolytic enzymes were upregulated and gluconeogenesis inhibited in cattle infected with T. congolense (135). It should be noted that all these analyses reflect aggregated outcomes, and there may be considerable variation in direction of metabolic changes between cell types and tissues. In contrast, no significant associations between glucose and disease severity indicators were observed in the CSF of T. b. rhodesiense-infected patients (241). Glucose consumption may directly reflect T. brucei metabolic needs since glycolysis is the major catabolic pathway for ATP production in bloodstream forms of the parasite (242).

Fig 4.

Glycolysis and TCA cycle. Glucose, the central fuel of glycolysis (yellow), is necessary to generate robust pro-inflammatory responses. Glucose-6 phosphate (P) can be diverted to the pentose phosphate shunt (blue). The end product of glycolysis, pyruvate, can be converted into acetyl-CoA to power the TCA cycle (gray). B cell generation of antibodies, TH2 immunity, and ROS generation are associated with TCA cycle activity. Kinetoplastids (A–C) and. Apicomplexans (D and E) may benefit from the activation of glycolysis. However, the enrichment or depletion of glycolytic metabolites is dependent on tissue identity and time of infection. Connections from the pentose phosphate pathway back toward glycolytic intermediates are not displayed, for simplicity. Figure created with BioRender.com.

Alternatively, glucose demand may reflect the importance of glycolysis for immune cell functions. Glycolysis is essential for TH17 cell survival in hypoxic environments (243). Glucose uptake is also critical for effector T cell expansion (244) and M1 macrophages (245) but not B cell activity (Fig. 4) (246). Activated T cells may, thus, also be a source of the observed elevated bloodstream lactate (245). T. brucei can increase glycolysis in neutrophils, possibly also leading to the observed glucose and lactate changes (247). Glucose uptake is necessary for production of NO and reactive oxygen species (ROS) in T. cruzi-infected macrophages (248).

During acute CD, glucose levels were decreased in plasma, while cardiac tissue glucose was increased in T. cruzi-infected mice (Fig. 4B) (107). Elevated cardiac glucose may be fueling glycolysis, as several glycolytic intermediates such as glucose-6-phosphate and fructose-6-phosphate were also increased in the T. cruzi-infected heart (107). This increased glycolysis improves T. cruzi invasion as well as replication (249), indicating that T. cruzi amastigotes modulate host energy metabolism to benefit their life cycle. Increased glycolysis was also associated with enrichment of interferon and TLR signaling signatures in T. cruzi-infected cardiomyocytes (249). Pyruvate and lactate were also increased in the heart during acute T. cruzi experimental infection (107). This may reflect the observed decrease of pyruvate dehydrogenase transcripts and activity in the acutely infected mouse heart, shunting products of glycolysis toward lactate rather than the TCA cycle (190). Activation of glycolysis is essential for production of NO, IL-1β, and IL-6 in monocytes/macrophages of chronic CD patients (250). Also, the pentose phosphate pathway is critical for production of NO and ROS in T. cruzi-infected macrophages stimulated with IFN-γ (248). TNF-ɑ signaling enhanced glycolysis and lactate production in vitro in muscle cell lines (251). TGFβ is considered one of the crucial drivers of collagen production in the heart during chronic CD (252). Also, it has been reported that TGFβ promotes glycolysis in primary human lung fibroblasts (253).

In contrast to CD, serum glucose levels were elevated in active visceral leishmaniasis patients, likely reflecting lower carbohydrate catabolism as lactate and malate were decreased (Fig. 4). Supporting these results, lower transcripts for the GLUT1 glucose transporter and for hexokinase 1, the initial enzyme in glycolysis, were observed in the spleen of C57BL/6 mice infected with L. donovani, 14 and 28 days post infection (254). However, a few transcripts for glycolytic enzymes were elevated in the spleen of hamsters infected with L. donovani 28 days post infection (103) and hexokinase-1 and glucose transporter transcripts were elevated in the spleen of BALB/c mice infected with L. donovani, 14 and 28 days post infection (254). Inhibiting glycolysis in mice infected with L. donovani increased liver and spleen parasite burden and decreased signatures of neutrophil activation (255). Thus, the impaired glycolysis in active visceral leishmaniasis patients (100) may be causally contributing to disease pathogenesis.

Apicomplexans

Human population studies have established that malaria-endemic regions have a high frequency of glycolytic enzyme variants, including the genes encoding glucose-6-phosphate dehydrogenase (G6PD) and pyruvate kinase (256–258). The stabilization of these alleles indicates that they likely confer a protective advantage to the host during Plasmodium infection, potentially by modulating erythrocyte metabolism. It is notable that none of these variants correlates with better survival in human malaria. However, this may be due to parallel parasite adaptations that have accumulated over time (201, 259). A not mutually exclusive model is that these metabolic enzyme polymorphisms have divergent roles depending on cell type and the stage of infection. For example, G6PD deficiency has been associated with host protection during cerebral malaria, which reduced the glycolytic burst from immune cell activation in the brain but exacerbated anemia related to erythrocyte function (260, 261). This hypothesis is further supported by the result that mice with severe P. berghei infection are depleted in the circulating glucose metabolism intermediates fructose bisphosphate, bisphosphoglycerate, and lactate (Fig. 4D) (262).

Notably, Plasmodium relies on glycolysis for asexual replication in erythrocytes (263). 2-DG treatment blocks male gametogenesis (male gametes lack both mitochondria and apicoplast), whereas female gametocytes retain a mitochondrion, allowing them to use glutamine as an ATP source (109). Glycolytic flux analysis has also shown that mice infected with P. berghei ANKA have increased levels of lactate in the liver and brain, particularly during cerebral malaria. Using labeled glucose, mice infected with P. berghei ANKA were shown to increase production of lactate and glutamate in the liver and brain, indicating increased flux through glycolysis and the TCA cycle when parasitemia was less than 1% (264).

Increased glycolytic flux is a well-described requirement for immune cell function and generation of the inflammatory response (9). However, there is evidence that Cryptosporidium, which has limited metabolic capabilities (60), may benefit by promoting host glycolytic metabolism for intracellular growth. C. parvum-infected human colorectal adenocarcinoma (HCT-8) cells increased glucose uptake and lactate release (265, 266). Mice infected with C. parvum or injected with soluble tachyzoite antigen reduced expression of gluconeogenesis enzymes and lactase in the ileum (68). Pharmacological inhibition of glycolysis in C. parvum infection significantly reduced intracellular parasite replication (108) although it is notable that this effect was not observed in for P. chabaudi infection (155). Karpe et al. found that the small and large intestines of C. parvum-infected mice transcriptionally upregulate the glycolytic enzymes phosphoglycerate kinase and glycogen phosphorylase compared to uninfected C57BL/6J mice, suggesting that host glycolysis may also be upregulated during in vivo infection. Currently, it is not clear if this is driven by the metabolic demand of the immune response, or a parasite manipulation strategy (267).

Although T. gondii has its own glycolytic machinery (two pyruvate kinases and AMPK), they are selectively upregulated during parasite egress and AMPK is dephosphorylated after cell entry, suggesting that glycolysis is used for short-term, extracellular energy demand when metabolite scavenging is not possible (268). This is also consistent with a model where T. gondii acquires metabolites primarily through intracellular rather than extracellular scavenging (268, 269). Proteomic analysis indicated that intracellular parasites are enriched for TCA cycle enzymes compared to extracellular parasites where glycolytic enzymes were more abundantly expressed. In parallel, host cells had an increase in NFkβ-dependent cytokines and glycolysis components (270). In partial alignment with this report, dual metabolic profiling of host and T. gondii metabolites in infected fibroblasts showed that both organisms increased metabolic intermediates in the TCA cycle and the pentose phosphate pathway (54). A potential limitation of this study is that it is not possible to differentiate between host and parasite TCA cycle metabolites which are structurally identical. However, paired transcriptional analysis indicated that host enzymes were largely unchanged, linking metabolite shifts to the parasite. The pentose phosphate pathway metabolizes glucose-6-phosphate and is a major source of NADPH for biosynthetic reactions, and pentoses as precursors for nucleotide biosynthesis. Deleting T. gondii ribulose-5-phosphate isomerase decreased glucose flux, formation of secretory organelles, and parasite growth proteins, which may explain why the parasite has evolved metabolic enzymes that are specialized to favor carbon cycling through the pentose phosphate pathway (271).

TCA cycle

Acetyl-CoA generated from lipid and carbohydrate oxidation feeds into the TCA cycle, leading to the release of substrate-derived carbons as CO₂, formation of reducing equivalents (NADH and FADH₂) that will be used to generate ATP during oxidative phosphorylation, and formation of GTP (Fig. 4). IL-4, associated with TH2 responses, promotes TCA cycle, whereas IFNγ, the canonical TH1/M1 cytokine, blocks TCA cycle and promotes glycolysis (170). TCA cycle transcripts were elevated in the blood of cattle infected with T. congolense (135). Citrate was decreased in the plasma of T. b. rhodesiense-infected humans at timepoints of active infection (Fig. 4A) (97). Citrate re-direction to fatty acid synthesis is important for B cell antibody production, prostaglandin, and reactive oxygen species synthesis (245). Similarly, the level of fumarate was decreased in the heart during acute CD in mice (Fig. 4B) (107), consistent with the observation that T. cruzi-infected cardiomyocytes are enriched for succinate, not downstream intermediates (249). Succinate promotes macrophage IL-1β production and reactive oxygen species generation. Thus, elevated succinate may be fueling antiparasitic responses and driving collateral tissue damage (245). Although T. gondii infection increases accumulation of TCA cycle intermediates in fibroblasts, during chronic infection, TCA cycle intermediates downstream of alpha-ketoglutarate were significantly depressed in the sera of infected mice compared to uninfected littermates (Fig. 4E) (54, 195).

FUTURE DIRECTIONS

Therapeutic potential of metabolic manipulation

Druggable metabolic targets to limit parasite survival

As our understanding of parasite metabolism and host metabolite scavenging has evolved, there has been interest in targeting these pathways for therapeutic intervention. For example, Toxoplasma uses two acyl-CoA acyltransferases to scavenge acetyl-CoA as well as an acyl-COA:diacylglycerol acyltransferase to import cholesterol esters from the host, which are essential for parasite survival (272, 273). Lipolysis of cholesterol esters and esterified lipids was recently shown to be regulated by a class of serine hydrolases which were previously thought to be depalmitoylating enzymes (274). Deleting this pathway led to toxic accumulation of cholesterol in the parasite, so targeting serine hydrolases may be a novel pathway to limit parasite growth in a manner that is less likely to have negative consequences for the host than targeting cholesterol import machinery or bioavailability directly. Several Plasmodium enzymes are sufficiently distinct from mammalian machinery that they may be druggable therapeutic targets for parasite clearance (78, 79). For example, the apicoplast is an organelle distantly related to the red algae plastid with metabolic and signaling pathways lacking a counterpart in humans. Despite having structurally divergent enzymes in the type II fatty acid synthesis pathway, these targets were shown to be dispensable for blood stage Plasmodium survival. However, recent studies have shown that P. falciparum requires this pathway for development in mosquitoes, suggesting that targeting apicoplast biology in vectors may be a novel intervention strategy (78, 79). These data also suggest that there may be a previously overlooked role for type II fatty acid synthesis pathway in liver stage P. falciparum, similar to the biology that was previously demonstrated in P. berghei during murine infection (275). Moreover, discovery that the synthesis of isopentenyl pyrophosphate by the apicoplast was necessary for parasite viability has opened the possibility of designing apicoplast-targeting therapeutic strategies for blood stage malaria.

In contrast, attempts have already been made to target metabolites that are used by both host and parasite to blunt parasite growth and persistence. Purine and purine-nucleoside analogues can act as potential trypanosomacidal and leishmanicidal compounds due to their antiproliferative effects. For example, 3′-deoxy-7-deazaadenosine has potent activity against T. cruzi in vitro and in vivo (276). The purine analogues azathioprine and 6-mercaptopurine displayed inhibitory effects against intracellular Leishmania amastigotes (277). Other purine analogs are active against T. cruzi and T. brucei (278, 279), though effects of allopurinol (hypoxanthine analog) in humans were more mitigated, possibly due to availability of hypoxanthine itself (280, 281).

Anorexia, a classic sickness behavior regulated by inflammatory cytokines, protects the host during sepsis by limiting glucose availability (282). Caloric restriction in mice is sufficient to limit Plasmodium load and dissemination to the brain; this was dependent on Plasmodium expression of the kinase KIN, which appears to function as a parasite nutrient sensor (283). Similar observations have been made in malaria patients brought to clinic during famines, who exhibited a spike in parasitemia shortly after re-feeding (284). For these reasons, glycolysis inhibitors have been proposed for malaria, particularly in chronic infection where a glycolytic inflammation can promote tissue pathology. However, the brain relies exclusively on glucose and ketone metabolism, so neurotoxic stress has been a concern for glycolytic therapies. To circumvent this problem, an approach to selectively inhibit glycolysis in erythrocytes has been proposed based on this cell type’s unusually high reliance on the γ-enolase (ENO2) for glycolytic function. Neurons express both ɑ-enolase (ENO1) and ENO2 which may allow them to adapt to ENO2 inhibition (285–287). Similar approaches could be considered to target cutaneous leishmaniasis based on the unique metabolic profile of the skin relative to other tissues in the body (161).

Major drug development efforts in Leishmania and T. cruzi have focused on targeting parasite ergosterol metabolism. However, clinical trials of inhibitors of ergosterol biosynthesis were unsuccessful in chronic Chagas disease, likely due to the fact that parasite proliferation (and thus ergosterol need) is lower at this disease stage (288, 289). Because these compounds are highly efficacious in vitro, other chemical scaffolds with the same target tend to be re-discovered in high throughput screening assays, hampering discovery efforts (290). Amphotericin B, which directly binds to membrane ergosterol, is successfully used to treat visceral leishmaniasis (291). It has not been successfully implemented for Chagas disease because the preferential accumulation of amphotericin B in the lungs, liver, and spleen make it well suited for L. donovani tropism but not for T. cruzi tissue tropism (292). These failures highlight the key need to consider differences in parasite metabolism between in vitro systems, animal models and humans, and also the importance of targeting slow-replicative or dormant stages of the parasite (4), rather than just actively-replicating cells. Tissue drug distribution, and its relation to parasite tropism, are further important considerations. A further, likely linked, limitation is that many of these studies have focused on drug repurposing. While this may reduce drug development costs, it means that these compounds were designed for different indications and may not be the best option for either parasite metabolic enzyme or parasite tissue tropism. A gap in funding for clinical trials and late-stage drug development [the so-called “valley of death” (293)] is particularly acute in these neglected diseases, meaning that many promising metabolism-targeting drugs do not get further developed, with the possible exception of studies in malaria.

Druggable metabolic pathways to protect host from pathology associated with parasite infection

“Disease tolerance” mechanisms limit bystander tissue damage caused by infection and inflammation, protecting the host without directly restricting pathogens (294). Immune-metabolic targets are a promising new avenue to manipulate disease tolerance programs without negatively impacting antiparasitic immunity (294, 295). For example, TGFβ-induced glycolysis in fibroblasts is essential for myofibroblast differentiation and collagen production (253). As TGFβ signaling plays an important role in the regulation of cardiac fibrosis and function during chronic CD, targeting the tissue repair/TH2 metabolic program associated with TGFβ production may limit cardiac damage due to fibrosis without negatively impacting parasite restriction (252). IL-1R signaling protects adipose tissues from necrosis during acute T. gondii infection, without altering parasite load. However, chronic activation of this pro-tolerogenic pathway resulted in fibrosis of metabolic tissues associated with cachexia (172, 296). Targeting tolerance pathways have the potential to be far safer than general immunosuppressants like cortisone due to their ability to leave pathogen-restricting immune mechanisms intact. However, there may still be maladaptive consequences of sustained reliance on tolerance programs that should be carefully considered.

An alternative approach is to address the metabolic pathways associated with tissue dysfunction during infection independent of immunometabolism. For example, supplementing carnitine during acute T. cruzi infection was sufficient to rescue mouse survival, reverse metabolic perturbations in the cardiac tissue, and improve cardiac function, without altering parasite load or tissue immune responses (191). Metformin treatment, with pleiotropic metabolic and antioxidant effects, improved cardiac oxidative damage and cardiac function, without changes in cardiac parasite burden in a chronic mouse model of T. cruzi infection (297). Such approaches are particularly important, given that late-stage antiparasitic treatment cannot improve Chagas disease patient outcomes (298) and is insufficient to restore metabolism (163, 299). In contrast, treatment regimens that address infection-induced immune dysregulation lead to improved cardiac metabolism, even with persisting parasites (163). During anorexia, host metabolism is skewed toward lipid oxidation. However, in malaria, this is complicated by parasite lysis of the red blood cell leading to anemia and limited oxygen availability. In contrast to the model discussed above, treating P. chabaudi-infect mice with glucose enhanced survival during the anorexic phase of the disease, whereas the glycolytic inhibitor 2-deoxyglucose (2-DG) reduced survival (155). Interestingly, 2-DG treatment did not directly impair parasite restriction, as peak parasite load remained the same as untreated animals. However, the duration of parasitemia, weight loss, and temperature exhibited a delayed return to homeostasis (155). These data link metabolic regulation to disease tolerance adaptations that allow the host to function during inflammatory stress and provide causal insight into the role of the targeted metabolic pathways in disease pathogenesis. Going forward, a challenge will be personalization of these interventions based on the pleiotropic nature of “normal” metabolism between individuals, as well as cross-tissue variability in metabolic needs and metabolic responses.

Promoting metabolic health in parasite infection by targeting commensal microbiota

Commensal microbiota play a critical role regulating immune development, parasite-specific immunity, and metabolic resources of the host and may be an additional node for therapeutic regulation (300, 301). T. gondii leads to polarization of the commensal community toward Gram-negative pathobiont species of bacteria in mice (302–305). C57BL/6J mice infected with C. parvum-infected have elevated D-alanine, D-norleucine, and D-proline in the small intestine, which is associated with fecal abundance of D-amino acids-producing microbes Lactobacillus, Lachnoclostridium, and Lachnospiraceae in the small intestine (267).

Short-chain fatty acids (SCFAs), generated by bacterial fermentation of polysaccharides in the gut, are critical to maintain systemic immune homeostasis through direct and indirect signaling to immune cells (306, 307). The cecum and colon of C. parvum-infected C57BL/6J mice harbored high levels of Blautia and Roseburia bacterial families, which can produce acetate and butyrate (267). Deprivation of dietary fiber, a major precursor of SCFAs, results in a lower abundance of Bacteroidetes bacteria in the large intestine and increased susceptibility of the host to C. parvum and C. tyzzeri (308). Bacteroides thetaiotaomicron (Bacteroidetes phylum) was also decreased by acute T. cruzi infection in BALB/c mice (309). Likewise, phylum Verrucomicrobia, which includes the butyrate-producing genus Akkermansia, was reduced in cardiac Chagas disease patients compared to controls and to other forms of Chagas disease (310). SCFAs are necessary for the generation of regulatory T cells subsets in the gut, which prevent IL-17-associated ileitis, as reviewed extensively (311, 312). SCFA also have systemic effects; relevant to the context of Chagas disease, they can help mitochondrial function, prevent cardiac inflammation, and serve as a fuel for cardiac metabolism (313). Mice with cryptosporidiosis also showed increased small intestinal Coriobacteriaceae (267). Although the consequence of modulating these species is not clear in infection, Coriobacteriacease have been linked to the improvement of glucose metabolism in type 2 diabetes (314), production of secondary bile acids (315), and increased levels of hepatic triglycerides and circulating high-density lipoprotein cholesterol (HDL ) (316) in mice.

More broadly, T. cruzi infection persistently modifies gut-associated microbial communities (317), and these changes are not restored by antiparasitic treatment (318). Given the strong metabolic function of the microbiome (309), and the correlations between microbiome and metabolome changes during infection (317), these persistent community changes may be driving the post-treatment metabolic sequelae observed in the heart following T. cruzi infection and antiparasitic treatment (163) and may contribute to the worsening patient outcomes following late-stage benznidazole treatment (298). A particular link may be at the level of purine metabolism (309). Consequently, interventions that restore microbiome composition and metabolism may, in turn, also lead to improved cardiac function in Chagas disease and to improved patient outcomes even following late-stage treatment.

Lastly, the therapeutic potential of manipulating commensal biology is not limited to gastro-intestinal infection or orally infectious parasites. A 2022 study directly comparing intraperitoneal infection to per oral infection with two type II Toxoplasma bradyzoite cysts demonstrated that bypassing the gut still led to commensal dysbiosis (319). Commensal homeostasis are severely perturbed following infection with Leishmania major, and ability to restrict these parasites is extremely attenuated in germ-free mice (320, 321). Skin microbiota constituents also play an important role regulating the immune response to cutaneous leishmaniasis, which may be an additional means of site-directed therapeutic delivery (321, 322). In contrast, antibiotic treatment protected hamsters from severe visceral leishmaniasis symptoms, through a disease tolerance-associated mechanism in the absence of changes in parasite burden (102). However, this was not associated with metabolic restoration in the liver or spleen (101).

Technical limitations and directions for metabolic tool development

Unlike transcriptomics, metabolic techniques cannot directly reveal the source of the observed metabolites, unless backed by knowledge of unique metabolic pathways. In the case of chronic infection, low parasite biomass, combined with detection limits of analytical instrumentation, can help speculate a host origin for observed metabolite increases, while confounding the ability to resolve parasite metabolic signatures. In contrast, during acute, high-burden infection, metabolites observed in biofluids and theoretically produced by both host and parasite could come from either organism. A parasite origin of unique, non-host-associated metabolites can be determined by comparative analysis of samples from parasite monoculture and infection samples but fails to recapitulate the diversity and relative abundance of metabolites generated by parasite biology in vivo. Metabolic labeling prior to infection could help untangle their relative roles in acute infection scenarios. Alternatively, pharmacological inhibitors specific to parasite metabolism could be used although these often inhibit host cell biology in addition to the parasite, and membrane permeability must be carefully considered for intracellular parasites (apicomplexans, Leishmania and T. cruzi) and in the setting of central nervous system infection where the blood-brain barrier can prevent inhibitor accessibility. Nevertheless, their successful implementation is valuable to demonstrate causal association of the targeted pathway in disease.

It is very likely that tools for subcellular metabolomics and single cell metabolomics will be transformational when applied to understand host-parasite interactions. These include advances in single-cell Raman spectroscopy and secondary ion mass spectrometry, among many other new techniques (323–325). The ability to directly sample from individual intracellular parasites will enable significant insight into in vivo and in situ parasite metabolism, as well as incontrovertible delineation of metabolite origin from host vs parasite. Some of these insights can be achieved by dual host and parasite RNA-seq. However, transcriptomics can only indicate the presence of metabolic enzyme transcripts and do not reflect regulation of enzyme activity by post-translational modifications or substrate vs product availability. Likewise, although purification of intracellular parasites from infected cells has proven powerful [e.g., references (208, 326)], this method, unlike single-cell, single-parasite approaches, cannot resolve the variability between individual parasite cells. This is a major concern given observed differences in chromosome copy number (mosaic aneuploidy), replication rates, and drug sensitivity between individual parasite cells (4, 327, 328).

Spatial considerations are a critical yet under-appreciated factor. For logistical reasons, most metabolomics studies of apicomplexan or kinetoplastid infection (especially in human cohorts) have relied on biofluid analyses. However, this ignores the influence of the local microenvironment on nutrient availability and on immune responses. For example, hepatocyte biology is dependent on proximity to venous vs arterial vasculature (329). Spatial heterogeneity is, therefore, expected to lead to hyperlocal differences in parasite and host metabolism, and consequently on pathogenesis (161, 162, 191). Instead, biofluid sampling should be considered in the context of cross-organ or cross-cell communication, where small-molecule signals from one infected or damaged site can affect other locations. This is better understood in the context of viral infection, through triggering of type I interferon responses restricting viral infection of neighboring cells (330). In contrast, bystander effects in the context of parasitic infection have been under-investigated: most studies usually focus on the site of highest parasite load. However, Leishmania parasites can perturb the microbiome and metabolome at sites distal to the lesion (118, 321). Likewise, with cardiac and gastrointestinal T. cruzi infection, sites of highest metabolic perturbations differ from sites of highest parasite burden (191, 193). In vitro, metabolic changes are observed not just in parasite-containing cells but also in neighboring uninfected (bystander) cells (331). These studies will benefit from spatial metabolomics approaches, including single-cell metabolomic techniques that record spatial information, “chemical cartography” (systematic mapping of tissue metabolism by liquid chromatography-mass spectrometry), and high-resolution mass spectrometry imaging (331–334).

Most studies summarized in this review rely on steady-state, single-timepoint snapshots of metabolism. However, these studies fail to appreciate metabolic flux. As an example, acylcarnitine accumulation is commonly reported (101, 107, 140). Do these primarily reflect increased formation of acylcarnitines by carnitine palmitoyltransferase in response to increased fatty acids (335)? Or is this an indicator of blocked downstream fatty acid oxidation? Both production and breakdown of acylcarnitines could also be increased, albeit at different rates. A common analogy is that of a picture of cars on a busy highway: from the picture, you cannot tell if cars are moving or if you are observing a traffic jam (336). A mechanistic understanding of the metabolic pathways driving differences in relative metabolite abundance will require flux analysis with labeled metabolic substrates.

Our limited functional understanding of most metabolite signals is another complicating factor. This is particularly pervasive in lipidomics: for example, nearly 250,000 lipid species with unique masses have been assigned numerical identifiers, but frequently only 10% of lipids detected in a mass spectrometry run have a known function (337). Metabolomics and lipidomics are further hampered with regard to annotation, something that has been described as the “dark matter” of the metabolome (338). This refers to metabolite features that are detected but cannot be confidently assigned a structure. This issue is further amplified for infection by eukaryotic pathogens that contain an admixture of host-derived and parasite-synthesized metabolites that are far less studied than mammalian systems. However frustrating, current studies are focused on the tip of the iceberg and there is tremendous potential to capitalize on existing data sets by iteratively searching them against ever-improving libraries of known metabolites as well as through new data processing workflows, including new structure elucidation workflows (339–345).

Host-parasite metabolic interactions: final considerations

In laboratory science and, to a certain extent, in clinical health, we have historically treated infection based on the Anna Karenina principle: there are many ways for metabolism to be perturbed, as discussed here with parasite infection, but one way for metabolism to be normal. However, as we have enriched our understanding of the genetic and environmental control of metabolism, the concept of a normal metabolism is starting to fracture. Multiple healthy statuses may be possible and fluctuate based on dietary and environmental input. Variation in ‘healthy’ status may protect against specific infectious agents while predisposing host susceptibility toward other infections or etiologies of disease. An emerging implication would be that the healthy status depends on your current nutrition availability, immune responses, microbiome, likelihood of pathogen exposures, etc. A challenge, however, is that restoration may be difficult since so many metabolic and immune parameters need to go back to “normal.”

Normalcy may also depend on the time of day, as emerging findings are revealing that both host and pathogen metabolism are under circadian control (346–348). Few metabolomic studies report or standardize the time of day of sample collection. Future work should consider addressing this issue. One challenge may be the long timespan needed for multi-organ sample collection from animal models, where tissue harvests can span an entire day.

A further challenge to the definition of “normalcy” are post-infectious effects, where symptoms persist even after pathogen clearance. Indeed, we and others have shown that metabolism does not fully re-normalize even after T. cruzi parasite clearance (163, 299, 349). Given that an individual’s life history covers a sequence of colonization and infection events, each individual’s metabolic baseline and response to subsequent infection may be individually re-shaped by these prior events. However, mouse models of infection often focus on acute or early-chronic timepoints where parasite burden may still be high and lack this infectious history, and thus, this is often hard to model in the laboratory. The impact of this metabolic reshaping by life history may be one reason why mouse models with wild-like microbiomes, which may include or have included pathogens, are often better models of human disease (350, 351). Post-infectious or persistent metabolic alterations may also hamper clinical response to antiparasitic treatment. Addressing these issues may involve either targeting the persistently-affected metabolic pathways, or targeting the root cause of these metabolic alterations, which may include durable immune reshaping, perhaps through epigenetic changes (352). Metabolic studies of chronic infection, particularly at times of low parasite burden, recovery stages and post-infectious timepoints are, thus, critically needed across kinetoplastid and apicomplexan pathogens but are currently limited in number. Expanding research at late infection timepoints will be critical to help bring insight to clinical settings. A further intriguing question is whether post-infectious metabolic changes can present as transgenerational effects (353–355).

Lastly, Koch’s postulates and molecular Koch’s postulates are cornerstones of infectious disease research (356, 357). However, conceptually applying them to the study of pathogenic infection-induced metabolic changes is challenging. Metabolic impacts of infection usually present as an ensemble of often inter-dependent metabolite or metabolic pathway changes. Many of these effects may be small individually, though large cumulatively. As demonstrated throughout this review, infection-associated metabolic alterations are common across pathogens and may also present in non-infectious conditions (violating the first molecular postulate). While metabolic changes may, indeed, be driving disease pathogenesis, as discussed above, they may be dependent on individual life history. A further complication is the compensatory and plastic nature of metabolism, where inhibition of one pathway may be compensated by induction of other pathways, making implementation of the second and third molecular postulates challenging. In parallel, given our current broader understanding of the different outcomes of infection, especially disease tolerance mechanisms that promote survival independent of pathogen load (294), there is a need for new readouts of pathogenesis. There is, therefore, a need for a new dialog around assessment of causality in the context of complex metabolic traits during and after infection. Immunometabolic alterations should be at the center of this new framework of disease pathogenesis assessment.