Introduction

Upon activation, immune cells fundamentally change their metabolism, requiring an increased supply of nutrients, which is essential for their effective function. Additional nutrients are delivered to immune cells by redirecting metabolites away from other tissues, making the immune system metabolically privileged, or selfish within an organism. The concept of selfish immunity proposes insulin resistance as a mechanism by which activated immune cells ensure adequate nutrient supply (Straub 2014; Dolezal 2015; Krejčová et al. 2023). For example, we have recently shown that immune cells produce cytokines Upd2 and Upd3, which induce the production of the insulin signaling inhibitor Impl2 in muscle tissue via the Jak/Stat signaling pathway. Impl2 reduces the consumption of carbohydrates by muscle, making nutrients more available to immune cells (McMullen et al. 2024). Another mechanism by which immune cells gain privileged access to nutrients is through the release of adenosine. Extracellular adenosine signaling, through the adenosine receptor, reduces nutrient consumption by non-immune tissues and slows the development of the organism, again leaving more nutrients available for the immune system (Bajgar et al. 2015; Dolezal 2015; Bajgar and Dolezal 2018).

Adenosine signaling plays a variety of roles, especially in more complex organisms such as mammals. However, the release of adenosine as a signaling molecule from cells plays a universal role from the most primitive organisms, such as social bacteria, to humans (Newby 1984; Boison and Jarvis 2021). The release of adenosine is an indication of metabolic stress and signals to neighbouring cells to adjust their behaviour accordingly. For example, the bacterium Myxococcus xanthus responds to nutrient depletion by releasing adenosine, which initiates the aggregation of bacteria into fruiting bodies (Shimkets and Dworkin 1981). In mammals the heart muscles secrete adenosine during ischemia, which causes blood vessels to dilate to increase blood flow and delivery of oxygen and nutrients (Lloyd et al. 1988; Deussen 1989; Kroll, Deussen, and Sweet 1992; Reiss et al. 2019). Adenosine also plays an important role in the induction of sleep, torpor and hibernation, physiological states associated with energy conservation (Fredholm, Johansson, and Wang 2011; Boison and Jarvis 2021). Therefore, the use of extracellular adenosine by immune cells to suppress nutrient consumption in non-immune tissues, thereby conserving energy for the immune response, fits well with this evolutionarily conserved role of adenosine signaling.

Adenosine is produced intracellularly in two ways (Figure 1), by conversion from AMP by nucleotidases or in the S-adenosylmethionine (SAM) transmethylation pathway from S-adenosylhomocysteine (SAH) by S-adenosylhomocysteinase (Ahcy). The increase in AMP mainly reflects the lack of energy and metabolic stress. Cells use adenylate kinase to produce one ATP and one AMP from two ADP molecules. Increased AMP activates AMPK, which suppresses energy-consuming processes in the cell and activates energy-producing processes (Kloor and Osswald 2004; P. Dzeja and Terzic 2009). AMP is also converted by nucleotidases to adenosine, which is released from the cell via the equilibrative nucleoside transporter (ENT), where it signals the metabolic stress of the cell to surrounding cells and tissues via adenosine receptors (Dolezelova et al. 2007; Fenckova et al. 2011). This is typical of cells during ischemia when they are deprived of nutrients (Deussen 1989; Schädlich et al. 2023). Since activated AMPK suppresses the function of immune cells (Salminen, Hyttinen, and Kaarniranta 2011), we would expect that in the absence of nutrients and decreasing ATP levels, the accumulated AMP would be converted to adenosine and released from the cells, thus demanding a greater supply of nutrients.

Figure 1.

In the SAM transmethylation pathway, ATP and methionine combine to produce SAM, which is the major methyl group donor for most methylations in the cell. The SAM transmethylation pathway is one of the largest consumers of ATP when the adenosyl moiety of ATP is incorporated into SAM (i.e., it is not a conversion of ATP to ADP as in energy transfer). After the methyl group is transferred to the target molecule by methyltransferase, SAM becomes SAH, which must be rapidly converted to homocysteine and adenosine, or it will block further methylation. As methylation increases in the cell, more SAM is formed and at the same time SAH is cleared more rapidly, leading to an increase in the SAM:SAH ratio, known as the methylation index. SAH is converted to homocysteine and adenosine by Ahcy, which works in both directions; for the “SAM to homocysteine + adenosine” direction to prevail, homocysteine and adenosine must also be metabolized or excreted from the cell very rapidly (Lu 2000; Kloor and Osswald 2004; Yu et al. 2019; Roy et al. 2020; Vizán, Di Croce, and Aranda 2021).

Homocysteine can either be recycled back to methionine, or further metabolized in the transsulfuration pathway (source of cysteine, H₂S, and antioxidants) (Sbodio, Snyder, and Paul 2019). Adenosine can be catabolized to inosine by adenosine deaminase (Novakova and Dolezal 2011), or recycled to AMP by adenosine kinase and further to ATP by adenylate kinase, glycolysis or oxidative phosphorylation (R. J. Zeleznikar et al. 1990; Robert J. Zeleznikar, Dzeja, and Goldberg 1995; P. P. Dzeja, Chung, and Terzic 2007; P. Dzeja and Terzic 2009). The importance of adenosine recycling to AMP by adenosine kinase is indicated by the dependence of the methylation cycle on the concomitant function of this enzyme (Lecoq et al. 2001; Bjursell et al. 2011; Moffatt et al. 2002). Since every single methylation in the cell is associated with the formation of one adenosine molecule, the amount of adenosine accurately reflects the intensity of the SAM transmethylation pathway. In a sense, it reflects the activity of the cell, because the more processes the cell has to activate, the more new molecules it has to methylate.

We have previously shown that activated immune cells preferentially acquire nutrients during immune response by releasing adenosine, subsequently reducing nutrient uptake in non-immune tissues, thereby leading to a delay in development (Bajgar et al. 2015; Bajgar and Dolezal 2018). To elucidate the origin of adenosine in activated immune cells, we investigated the role of the SAM transmethylation pathway. Given its upregulation and critical function in mammalian immune cells (Lawson et al. 2012; Yu et al. 2019; Roy et al. 2020), we hypothesized that this pathway is also a major source of adenosine in Drosophila immune cells. Our results confirm that the SAM transmethylation pathway is indeed upregulated in Drosophila hemocytes upon parasitoid wasp infection. In addition, we show that this pathway is not only essential for the immune response, but also a source of extracellular adenosine that slows larval development. Moreover, not all adenosine is released from hemocytes, but it can be recycled back into ATP and thus re-enter the SAM transmethylation pathway.

Results

SAM transmethylation pathway accelerates in activated hemocytes

One of the important sources of adenosine in hemocytes is the SAM transmethylation pathway (Figure 2A). Putative methionine transporters, as well as all enzymes of this pathway, are strongly expressed in hemocytes, both in the resting and activated state (Figure 2B, Figure 2—figure supplement 1). Methionine can be transported from the hemolymph to the cytosol by Drosophila L type amino acid transporters orthologs mnd (CG3297; FBgn0002778), Jhl-21 (CG12317; FBgn0028425), CG1607 (FBgn0039844), sbm (CG9413; FBgn0030574), or gb (CG6070; FBgn0039487). The expression of mnd and gb is particularly high and is further increased during infection (Figure 2B). SAM synthetase (Sam-S; CG2674; EC 2.5.1.6; FBgn0005278) converts methionine and ATP into SAM. SAM is used for methylation by a wide range of methyltransferases (EC 2.1.1.-), or in polyamine synthesis, by SAM decarboxylase (SamDC; CG5029; EC 4.1.1.50; FBgn0019932), producing 5‘-methylthioadenosine (MTA). Methylation reactions produce SAH, a potent inhibitor of methyltransferases which must be quickly removed to keep the pathway running. Ahcy (Ahcy; CG11654; EC 3.3.1.1; FBgn0014455) hydrolyzes SAH into adenosine and homocysteine and shows the highest expression among SAM transmethylation pathway enzymes (Figure 2B). AhcyL1 (CG9977; FBgn0035371) and AhcyL2 (CG8956; FBgn0015011), which work as dominant-negative regulators of Ahcy (Parkhitko et al. 2016) have a lower expression and are downregulated during infection (Figure 2B). When adenosine and homocysteine accumulate, Ahcy changes the activity in favor of SAH synthesis. Thus, adenosine and homocysteine must be rapidly metabolized (Kloor and Osswald 2004). Homocysteine can be remethylated back to methionine, by betaine-homocysteine methyltransferase (Bhmt; CG10623; EC 2.1.1.5; FBgn0032727) or homocysteine S-methyltransferase (CG10621; EC 2.1.1.10; FBgn0032726). Alternatively, it is used as a substrate for Cystathionine β-synthase (Cbs; CG1753; EC 4.2.1.22; FBgn0031148) to produce cystathionine (CTH) in the transsulfuration pathway (Figure 2A). Overall, enzymes of the SAM transmethylation pathway, and its associated metabolic branches, are strongly expressed in hemocytes, however, there is no specific pattern of changes upon hemocyte activation during infection (Figure 2B).

Figure 2.

To analyze SAM transmethylation pathway metabolism, we used high pressure liquid chromatography coupled with tandem mass spectrometry (HPLC-MS/MS). In addition to detecting relative levels of metabolites, 13C-labeled methionine tracing was used to determine methionine usage in the SAM transmethylation pathway in hemocytes of infection and non-infected animals. Hemocytes were extracted by bleeding larvae at 20 hours post infection (hpi). Cells were incubated in media containing 0,33 µM ¹³C₅ methionine for 20 minutes. This ex vivo incubation was designed to resemble physiological conditions. In control larvae, the labeled intracellular methionine fraction reached almost 70% in resting hemocytes and over 80% in activated hemocytes after 20 min of ex vivo incubation, demonstrating a substantial uptake of methionine into hemocytes (Figure 2C). Much of this methionine is used for SAM production and methylation, as indicated by the 2-fold increase of SAM and SAH in infected compared to control animals (Figure 2D and E); demonstrating an upregulation of the SAM transmethylation pathway in activated hemocytes. This is further supported by a significantly higher methylation index (ratio of SAM:SAH; Figure 2H-J) in activated hemocytes.

SAH is cleaved into homocysteine and adenosine, however, we were unable to reliably detect homocysteine in our metabolomic analysis. Homocysteine can be recycled back to methionine, to enter another round of the SAM transmethylation pathway. In this scenario, fully labeled 13C5 methionine, as was used in this assay, would lead to the formation of methionine m+4. One labeled carbon is used in the methylation reaction while the unlabeled methyl group would be added during recycling (Figure 2A). However, the m+4 portion of the labeled methionine (m+4 and m+5) did not exceed the 3% threshold, set by the methionine labeling impurity (ratio “m+4/[m+4 + m+5]” in the medium was 3%; Figure 2F). This indicates that methionine used for the SAM transmethylation pathway in hemocytes is derived from hemolymph and homocysteine is not remethylated to methionine.

To investigate the role of the SAM transmethylation pathway in hemocytes, we knocked down Ahcy by hemocyte-specific RNAi (Figure 2—figure supplement 1 and 2). Ahcy RNAi leads to reduced 13C-labeled methionine uptake in resting hemocytes, although uptake is still increased upon infection (Figure 2C). Methionine incorporation into SAM is significantly lower in activated hemocytes (Figure 2D) and incorporation into SAH is decreased in both, quiescent and activated hemocytes (Figure 2E). These results imply that the SAM transmethylation pathway slows down when there is insufficient Ahcy available. This is reflected in metabolite levels, where SAH is heavily accumulated in Ahcy RNAi hemocytes (Figure 2I) and SAM and methionine levels are higher compared to controls (Figure 2G and H); indicating that they are not metabolized in the SAM transmethylation pathway. This is substantiated by a methylation index that is very low compared to controls, and does not increase upon infection (Figure 2J).

SAM can also be utilized for polyamine synthesis that, in addition to polyamines, produces MTA. Polyamines are essential for cell growth and proliferation and participate in various cellular processes (replication, gene expression, biomolecules modifications, etc.). They play a role in macrophage polarization and regulation (Latour, Gobert, and Wilson 2020; Xuan et al. 2023). There is significantly higher incorporation of 13C methionine into MTA in activated hemocytes (Figure 2K) with no changes in MTA level (Figure 2L). Ahcy RNAi hemocytes have increased MTA levels (Figure 2L), which can reflect MTA accumulation, due to decreased SAM transmethylation pathway activity. Decreased 13C methionine incorporation into MTA upon Ahcy knockdown by RNAi (Figure 2K) may be caused by dilution, as unlabeled MTA accumulate (Figure 2L) therefore, we are unable to determine the effect of Ahcy RNAi on polyamine synthesis.

The SAM transmethylation pathway is connected to the transsulfuration pathway, which produces cysteine and further antioxidants taurine, and glutathione (Sbodio, Snyder, and Paul 2019). The transsulfuration pathway is one of the ways to metabolically remove homocysteine produced by Ahcy when Cbs converts homocysteine into CTH. This is supported by notable 13C methionine incorporation into CTH (Figure 2M). Interestingly, there is a small, but significant, drop in the labeled CTH fraction in activated hemocytes (Figure 2M), however, this could be due to an overall increase in the level of CTH, including the unlabeled form (Figure 2N). As reported above, the SAM transmethylation pathway is accelerated in activated hemocytes, producing more adenosine, as well as more homocysteine, which must be removed to keep the pathway running. Cystathionine γ-lyase (Cth; CG5345; EC 4.4.1.1; FBgn0000566) usually converts CTH to cysteine. However, this enzyme can also use homocysteine for H₂S production, and was shown to be responsible for homocysteine removal under hyperhomocysteinemia, when Cbs is unresponsive (Singh et al. 2009). H₂S has many diverse roles in the organismal physiology, dependent on its source, localization, and concentration, and has already been connected with immune cells (Sbodio, Snyder, and Paul 2019; Dilek et al. 2020). Based on our transcriptomic data, Cbs expression in Drosophila larval hemocytes is rather low, and even decreases during infection. Interestingly, Cth is strongly expressed, at similar levels to Ahcy. Therefore, our data gives evidence to the clearance of homocysteine by Cth and H₂S production rather than via CTH and Cbs.

Knockdown of Ahcy leads to the accumulation of SAH and thus the production of homocysteine, the substrate for CTH, should be strongly inhibited. However, our data shows the relative levels of methionine, as well as the incorporation of 13C methionine into CTH in Ahcy RNAi hemocytes is equivalent to controls (Figure 2M and N). This suggests, there is an alternative connection of methionine with homocysteine and consequently CTH, in addition to Ahcy. Uncovering this alternative connection would be of interest, although it is beyond the scope of this current work.

In summary, activated hemocytes increase the uptake of methionine to ramp up the transmethylation pathway. SAM is not only a substrate for methylation, but is also used for polyamine synthesis. Homocysteine is not recycled back to methionine, but is further metabolized in the transsulfuration pathway. Interestingly, our data suggest an alternative link between methionine and CTH independent of Ahcy. Hemocyte-specific Ahcy-RNAi potently suppresses the SAM transmethylation pathway.

Figure 2-figure supplement 1.

Figure 2-figure supplement 2.

Ahcy produces adenosine as a systemic signal that influences the immune response

In our previous work, we showed that activated hemocytes release increased amount of adenosine (Bajgar et al. 2015; Bajgar and Dolezal 2018). This causes a systemic metabolic switch that slows the organism’s development, leaving energy for an effective immune response. Our results show that the SAM transmethylation pathway is enhanced in hemocytes after infection. This means that increased levels of adenosine is produced, as every methylation event results in the production of one adenosine molecule. At the same time, the cell must rapidly remove adenosine, along with homocysteine, to avoid reversing the direction of Ahcy enzyme activity, which would block methylation. This appears to be reflected in the significant reduction in intracellular adenosine levels after infection (Figure 3A). At least part of this reduction is likely due to the release of adenosine from the cell (via the equilibrative nucleoside transporter ENT2 (Bajgar et al. 2015)) since extracellular adenosine levels increase after infection (Figure 3B), as does the level of inosine, which is formed from adenosine by the action of the adenosine deaminase ADGF-A. We measured the extracellular adenosine production in hemocytes after 20 minutes of ex vivo incubation, when hemocytes were removed by centrifugation and adenosine and inosine were measured in the collected supernatants. We also chose to detect inosine because activated hemocytes produce the adenosine deaminase ADGF-A (Novakova and Dolezal 2011). Therefore, some adenosine is always converted to inosine during sample processing, even when the adenosine deaminase inhibitor EHNA (erythro-9-(2-hydroxy-3-nonly)adenine) is used, otherwise it is difficult to detect any adenosine.

Figure 3.

Hemocyte-specific Ahcy RNAi leads to a significant decrease in intracellular adenosine levels, confirming that a significant fraction of adenosine is generated by the SAM transmethylation pathway (Figure 3A). However, there is no increase in extracellular adenosine, or inosine levels upon Ahcy knockdown in hemocytes following infection (Figure 3B). This indicates that the SAM transmethylation pathway is responsible for the generation of this extracellular signal. This is further supported by the fact that silencing of Ahcy in hemocytes during infection leads to a significant reduction (by 6 hours) in developmental delay (Figure 3C), the hallmark of systemic effects of adenosine during infection. Similar to our previous work (Bajgar et al. 2015), this results in diminished host defense, which is associated with reduced production of lamellocytes that encapsulate the parasitoid (Figure 3D) and reduced survival (Figure 3E). Interestingly, the induction of Ahcy RNAi in hemocytes accelerates development by approximately 2 hours, even in the absence of infection (Figure 3C), a result similar to silencing ENT2 specifically in hemocytes (Bajgar et al. 2015). The normal activity of hemocytes, associated with methylation and adenosine release, appears to slow the overall development of the organism.

In summary, the SAM trabsmethylation pathway and Ahcy produce the majority of adenosine in Drosophila larval hemocytes and strongly participate in the production of extracellular adenosine as a systemic metabolic regulator. Knockdown of Ahcy in hemocytes reduces larval developmental delay associated with infection, demonstrating the lack in extracellular adenosine production. In addition, knockdown of Ahcy in hemocytes negatively impacts immune response.

Figure 5-figure supplement 1.

Activated hemocytes recycle ATP and SAM from adenosine

The above results demonstrate that the SAM transmethylation pathway is a major producer of adenosine in activated hemocytes. To maintain methylation, the cell must rapidly remove adenosine, some of which is released from the cell. This, of course, reduces the adenosyl pool of the cell; note that the SAM pathway is a major consumer of ATP when it combines methionine with ATP to form SAM (Figure 4A). Therefore, activated immune cells have active de novo purine production, as shown in mammalian macrophages and T lymphocytes (Yu et al. 2019; Roy et al. 2020). Our recent work showed increased pentose phosphate pathway activity is linked with nucleotide production in activated hemocytes (Kazek et al. 2024). However, this is very energy consuming and requires many steps, whereas the recycling of adenosine to AMP requires only one step and one ATP. Ahcy activity has indeed been linked to adenosine kinase activity, which produces AMP from Ado (Moffatt et al. 2002; Xu et al. 2017; Murugan et al. 2021). AMP can then be converted to ADP by adenylate kinase (P. Dzeja and Terzic 2009) and finally to ATP by glycolysis or OXPHOS, which can again enter the transmethylation pathway.

Figure 4.

The most striking expression change we detected in activated hemocytes was the increase of adenosine kinase Adk3 (Figure 4B and C) and cytosolic adenylate kinase Ak1 (Figure 4B and D). Another adenosine kinase, Adk2, which is not altered by infection, is also highly expressed in hemocytes, as is the mitochondrial adenylate kinase Ak2 (Figure 4 B-D). Activated hemocytes have a system for recycling adenosine back to ATP, which appears to be further enhanced during infection. The increase of cytosolic Ak1 indicates at least partial redirection of energy metabolism from mitochondria (oxidative phosphorylation) into cytosol (aerobic glycolysis) in activated immune cells (Kornberg 2020).

The increase in all levels of IMP, AMP, ADP and ATP (Figure 4E-H) indicates that de novo nucleotide production occurs in activated hemocytes. De novo purine synthesis may play many different roles in the cell, but an important one is the replenishment of ATP for SAM synthesis. In addition to de novo synthesis, our expression analysis suggests that recycling of adenosine to ATP is also involved. We tested this by adding labeled adenosine to resting and activated hemocytes ex vivo (Figure 4A). Since hemocytes express ENT2 (Figure 4B and Bajgar et al. 2015), at least some of the labeled adenosine enters the hemocytes and is converted to AMP and finally to ATP (Figure 4I and J). This demonstrates that hemocytes recycle adenosine to ATP, in addition to de novo purine synthesis. Both the AMP m+5 and ATP m+5 fractions are lower after infection (Figure 4I and J). This is likely due to dilution of the labeled fraction by a significant increase in newly synthesized, unlabeled AMP and ATP (Figure 4F and H). Activated hemocytes also release more adenosine from the cell, so it is possible that it is more difficult for labeled adenosine to enter the cell. Therefore, the extent of ATP recycling from adenosine in resting and activated hemocytes remains elusive.

Resting hemocytes show a diminished fraction of labeled SAM derived from labeled adenosine (Figure 4K). After infection, there is a significant increase in SAM labeling (Figure 4K), with approximately 13% of labeled ATP molecules being used for SAM synthesis (compare Figure 4J to Figure 4K), while the total level of SAM does not change (Figure 4L). Since ATP is labeled to greater extent in resting hemocytes, this suggests a compartmentalization of the enzymatic cascade physically linking SAM to adenosine recycling (Pedley and Benkovic 2017; Bar-Peled and Kory 2022). If any ATP was used for SAM production, then a higher fraction of labeled ATP in resting hemocytes should be reflected in higher labeled SAM fractions. A higer fraction of labeled SAM in activated hemocytes suggests a physical connection of SAM synthesis and ATP recycling from labeled adenosine. Increased recycling of adenosine to ATP would be characterized by more 13C entering SAM, despite the fraction of labeled ATP appearing lower due to the total amount of ATP produced by de novo synthesis.

In conclusion, both resting and activated hemocytes recycle adenosine to ATP. Activated hemocytes strongly upregulate the expression of Adk3 and cytosolic Ak1 and increase production of SAM from recycled adenosine.

Adenosine kinase and cytoplasmic adenylate kinase are important for SAM pathway and immune response

To determine the importance of adenosine kinase and adenylate kinase, which seem to be necessary for recycling of ATP from adenosine, we again used hemocyte-specific RNAi induced by Srp-Gal4Gal80. Adk3 RNAi reduces the infection-induced increase in expression almost 8 times (Figure 5—figure supplement 1). Adk3 expression is low in hemocytes of uninfected larvae (Figure 4B and C), thus we did not observe any effect of Adk3 RNAi on the AMP level in uninfected larvae (Figure 5A). However, there is significantly lower level of AMP in hemocytes from the infected larvae (Figure 5A). Adk2 is also strongly expressed both in resting and activated hemocytes (Figure 4C) and that is, most likely, why we still see an infection-induced increase in AMP upon Adk3 RNAi (Figure 5A). This increase is also due to de novo purine synthesis. Nevertheless, the increase in AMP in Adk3 RNAi activated hemocytes is significantly smaller than that of controls, indicating that Adk3 activity is important for adenosine conversion to AMP in activated hemocytes. Increase in methylation index is abrogated upon Adk3 RNAi (Figure 5B), showing that Adk3 activity is required for the infection-induced increase in SAM transmethylation pathway. The consequences of Adk3 knockdown phenocopy those of Ahcy knockdown animals, as development is significantly less delayed during infection (Figure 5C) and number of lamellocytes (Figure 5D), as well as survival (Figure 5E) are decreased. We observe similar effects upon cytoplasmic Ak1 knockdown in hemocytes (Figure 5—figure supplement 1), another enzyme upregulated upon infection, which is likely involved in adenosine recycling. The methylation index is abrogated (Figure 5B), development is significantly less delayed during infection (Figure 5F), number of lamellocytes is decreased (Figure 5G) and survival is lower (Figure 5H) upon Ak1 RNAi mediated knockdown. In summary, recycling of adenosine to AMP and potentially to ATP and SAM, using adenosine kinase and cytoplasmic adenylate kinase, appears to be important for activated hemocytes to increase the SAM transmethylation pathway and for effective immune response.

Figure 5.

Discussion

We have previously shown that adenosine, released from Drosophila hemocytes, is required for systemic metabolic switch and to slow development so that there are enough nutrients for an effective immune response (Bajgar et al. 2015). Here, we show that this signal is generated in hemocytes by the SAM transmethylation pathway, which is enhanced upon infection and is required for the immune response. This novel observation links SAM transmethylation in immune cells to the production of adenosine as a systemic metabolic regulator. In addition, we show that adenosine is not only released from hemocytes but is also recycled to ATP and SAM, mediated by Adk3 and the Ak1, which are necessary for the increase in SAM transmethylation during infection, and therefore effective immune response.

The SAM transmethylation pathway is part of the one-carbon metabolism interconnecting the metabolism of serine, glycine, folates, and methionine, supporting de novo synthesis of purines and thymidine, and maintaining redox balance. The SAM transmethylation pathway itself controls antioxidant synthesis, sulfur metabolism, polyamine synthesis, and epigenetic programming via methylation (Ducker and Rabinowitz 2017; Vizán, Di Croce, and Aranda 2021). Therefore, Ahcy activity, removing SAH and producing adenosine and homocysteine, is essential, not only for keeping proper function of diverse methyltransferases, but also for a wide range of other essential metabolic processes inside the cell.

There is increasing evidence showing the crucial role of the SAM transmethylation pathway in immune cells (Lawson et al. 2012; Klein Geltink and Pearce 2019). It was shown as early as the 80s that peripheral blood mononuclear cells in a “resting” state consume approximately three to five times more SAM for methylation than other, nonhepatic, tissues. Once activated, their SAM consumption is further increased. High SAM utilization was also observed in cultured growing lymphoblasts (German, Bloch, and Kredich 1983). The proper function of the SAM transmethylation pathway and efficient removal of SAH by Ahcy is necessary for chemotaxis in human monocytes, and neutrophils and for lymphocyte effector function (Pike and DeMeester 1988; Zimmerman, Wolberg, and Duncan 1978; Pike, Kredich, and Snyderman 1978). More recently, stable isotope labeling techniques have allowed deeper insight into metabolic set up of activated immune cells. It has been shown that LPS-activated macrophages synergistically increase glycolysis, one carbon metabolism, serine synthesis pathway, pentose phosphate pathway, de novo purine synthesis, and SAM transmethylation pathway activity, to induce, and maintain a proinflammatory phenotype. Interestingly, the homocysteine produced by these macrophages is not recycled back into methionine, but SAM synthesis is likely fed exclusively with exogenous methionine transported into the cells (Yu et al. 2019). Roy et al. (2020) demonstrated the metabolic changes that occur in activated CD4+ and CD8+ T lymphocytes. They identified methionine as a crucial metabolite in driving T cells proinflammatory phenotype via epigenetic reprogramming and demonstrated dietary methionine restriction as an effective intervention for autoimmune disease. Activated T lymphocytes, like macrophages, upregulate glycolysis, serine, glycine, and one-carbon metabolism including the SAM transmethylation pathway. These studies indicated that the folate cycle and the SAM transmethylation pathway are uncoupled and homocysteine is not remethylated. Importantly, SAM is synthesized from exogenous methionine and ATP, predominantly produced by de novo purine synthesis. Here we show that in Drosophila immune cells, SAM is also synthesized mostly from exogenous methionine, and homocysteine is not remethylated back to methionine, but rather used in transsulfuration pathway. Interestingly, our results also suggest an alternative means of CTH production from methionine, independent of Ahcy, although this is not further explored here. Activated hemocytes also increase de novo purine synthesis, supported by increased pentose phosphate pathway activity (Kazek et al. 2024).

Ahcy was shown to play an important role in adenosine production in heart muscle cells, under normoxic conditions (Lloyd et al. 1988; Deussen 1989). Adenosine produced by Ahcy contributes to viability of hepatocellular carcinoma cells (HepG2), since Ahcy knockdown followed by adenosine depletion leads to activation of the DNA damage response, cell cycle arrest, a decreased proliferation rate and DNA damage (Belužić et al. 2018). However, the adenosine formed in the SAM transmethylation pathway is usually referred to in studies as a by-product that needs to be removed quickly to avoid Ahcy reverse function, which would block further methylation. Here, we show that adenosine formed in the SAM transmethylation pathway is not a mere by-product but can have a significant impact on the immune response and the physiology of the whole organism. With the increase of SAM transmethylation pathway activity in activated hemocytes, more adenosine is produced. As intracellularly adenosine decreases, it is released to the extracellular space; this leads to suppression of metabolism in non-immune tissues, and consequently a development delay. This metabolic suppression is crucial for redirecting nutrients towards the immune system and mounting an effective immune response. Knockdown of Ahcy by RNAi suppresses the infection-induced increase in extracellular adenosine, as well as the delay in development, demonstrating that this important systemic signal is indeed generated by the SAM transmethylation pathway. It is important to add that while adenosine-mediated metabolic suppression and developmental delay are important for an effective immune response, it is more likely that the ineffective immune response in Ahcy RNAi animals is due to suppression of the SAM transmethylation pathway itself in hemocytes, which is required for their function. At the same time, we do not want to suggest that SAM transmethylation in hemocytes is the only source of adenosine during infection. Ectoenzymes are an important source of extracellular adenosine that convert extracellular ATP to ADP, AMP and adenosine. This cascade is an important immunomodulator in mammals (Antonioli, Pacher, and Haskó 2022). Drosophila also has this cascade of ectoenzymes (Fenckova et al. 2011), which appears not to have an immunomodulatory role, likely due to more rudimentary adenosine signaling (Dolezelova et al., 2007), but could still serve as a source of extracellular adenosine. Nevertheless, the absence of extracellular adenosine increase and reduced developmental delay upon hemocyte-specific knockdown of Ahcy clearly demonstrates the essential role of the SAM transmethylation pathway in generating of adenosine as a systemic signal.

After ATP, SAM is the most used substrate in the cell, with up to half of the daily methionine intake in humans used for SAM synthesis, the vast majority of which is used for methylation (Lu 2000). If all the adenosine produced by Ahcy was released to hemolymph, hemocytes would likely lose a huge amount of ATP due to SAM synthetase utilizing the adenosyl group of ATP and methionine to create SAM. Increased ATP consumption by the accelerating the SAM transmethylation pathway must be supported by de novo purine synthesis. Previous work (Kazek et al. 2024) as well as results presented here, show increased de novo purine synthesis in activated immune cells. However, this consumes a considerable amount of energy (6 ATP for one IMP molecule) and takes eleven steps to produce IMP from the initial ribose-5-phosphate. Thus, it is more efficient to recycle at least some ATP from the adenosine generated in the SAM transmethylation pathway via coordinated action of adenosine kinase, cytosolic adenylate kinase, and glycolysis. This is in line with ATP synthesis from adenosine via AMP and ADP shown to be significantly increased in activated rat peritoneal macrophages (Barankiewicz and Cohen 1985).

Adenosine kinase has been shown to be coupled with Ahcy to effectively remove adenosine and keep the SAM transmethylation pathway running (Moffatt et al. 2002; Xu et al. 2017; Murugan et al. 2021). Subsequently, the AMP produced can be utilized by the adenylate kinase catalyzing phosphoryl exchange reaction AMP + ATP ↔ 2 ADP. Adenylate kinase is crucial for cellular energy homeostasis, by regulating nucleotides ratio, AMP production and direction to diverse metabolic sensors (e.g. AMPK). In addition, it is responsible for the transfer of ATP energy rich phosphoryl between energy producing and energy utilizing sites by a sequence of repeated reactions, creating an effective energy transfer shuttle across the cell (R. J. Zeleznikar et al. 1990; Robert J. Zeleznikar, Dzeja, and Goldberg 1995; P. P. Dzeja, Chung, and Terzic 2007; P. Dzeja and Terzic 2009). Our hemocyte RNAseq and qPCR data show that Drosophila Adk3 and Ak1 are expressed at low levels in quiescent hemocytes, but their expression increases dramatically during infection. Experiments using 13C labeled adenosine show its incorporation into AMP, ADP, ATP, and especially into SAM, in which the incorporation is significantly increased upon hemocyte activation. These results connect adenosine recycling with the SAM transmethylation pathway. Without infection, the methylation index of Adk3 and Ak1 RNAi in hemocytes is the same as in control cells. However, upon infection, there is no increase in the methylation index in Adk3 and Ak1 RNAi cells. When the SAM transmethylation pathway is not accelerated, Ahcy does not produce more adenosine, and consequently the adenosine-induced developmental delay during infection is reduced in Adk3 and Ak1 RNAi, whereas without infection, RNAi larvae develop at the same rate as controls. Our results therefore demonstrate a novel functional connection of adenosine recycling to ATP with SAM transmethylation pathway. At least under energy demanding conditions, as during the activation of immune cells, Adk and cytosolic Ak activities appear to support SAM transmethylation pathway activity.

Similar to Ahcy RNAi, Adk3 and Ak1 RNAi in hemocytes also negatively affects immune response efficiency. We see a significant decrease in lamellocytes number in our infected knockdown larvae, resulting in decreased survival. It is important to note, that while Ak1 RNAi most likely also affects de novo purine synthesis, the Adk3 RNAi effects we observed are due to impairments in adenosine recycling. Connecting the SAM transmethylation pathway-driven production of adenosine with its recycling via Adk3 and Ak1 leads us to following hypothesis (Figure 6):

Figure 6.

Most cellular processes involve methylation, therefore the SAM transmethylation pathway reflects the activity of the cell. While a wide variety of molecules are methylated, adenosine is the universal product of each individual methylation. Thus, adenosine production reflects the sum of all methylations (where SAM has been used) and therefore the overall activity of the cell. Adenosine must be removed quickly as to not block the SAM transmethylation pathway. The preferred fate is to recycle adenosine to AMP and eventually ATP so that the cell does not lose its adenosyl pool. However, this requires nutrients, adenosine kinase requires ATP, adenylate kinase produces ADP from AMP and ATP and the resulting ADP is recycled to ATP in glycolysis, consuming glucose. If the cell’s nutrient supply is inadequate, AMP begins to accumulate, due to the opposite action of adenylate kinase, preventing adenosine from being recycled to AMP by adenosine kinase. In this case, the accumulating adenosine is pushed out of the cell via ENT, where it becomes a signaling molecule, informing surrounding tissues of the need for more nutrients. We therefore look at adenosine as a sensitive sensor of the balance between cell activity (reflected in methylation) and nutrient supply - the better the balance, the less adenosine is released and the more is recycled.

To experimentally test this hypothesis will require extremely well controlled conditions, as offered for example by uniform cell culture in precisely controlled media, and delicate manipulation of the important players. The proposed role of adenosine as a sensitive sensor between cell activity and nutrient supply might have far-reaching implications. The production of adenosine by nucleotidase as the ADP/ATP ratio rises is known, for example, to generate a signal in ischemic tissue to inform surrounding tissues of metabolic stress (e.g., vasodilation to increase blood flow). However, the generation of such a signal as a result of an imbalance between cell activity and nutrient supply is a novel and exciting hypothesis.

The fate of adenosine is determined by biochemical properties and the expression of enzymes and transporters. The Km values of Drosophila Adk2 and Adk3 are not known (Fleischmannova et al. 2012; Stenesen et al. 2013), but it is possible that they differ, and that increased expression of Adk3 specifically during infection is important in readjusting the balances between AMP and adenosine, which are determined by the properties of nucleotidases and adenosine kinases (Boison 2013). The expression levels and properties of nucleoside transporters and adenosine deaminase also play important roles. The specific setup of this whole system in certain cell types under certain conditions determines what happens with adenosine (Dulla and Masino 2013; Dale 2021). Studying this setup is essential to understanding the role of this system in specific situations, such as those in which adenosine may be generated by imbalances in activity and energy supply. For example, does it play a role in the early activation of immune cells, regulating nutrient supply to their site of activation? In addition, in the tumor microenvironment, adenosine inhibits infiltrating immune cells (Boison and Yegutkin 2019). Until now adenosine ectoenzymes have been mainly studied in this respect. However, the SAM transmethylation pathway is very active in tumor cells (Vizán, Di Croce, and Aranda 2021) - can tumors regulate their environment by releasing adenosine, which is produced according to our hypothetical mechanism?

Methods

Fly strains, cultivation, and parazitoid wasp infection

Drosophila melanogaster strain w¹¹¹⁸ (FBal0018186) in Canton S genetic background (FBst0064349) was used as a control line unless otherwise stated. Strains UAS-Ahcy-RNAi Ahcy^HMS05799 (FBti0186788), UAS-Adk3-RNAi Adk3^HMC06354 (FBti0185498), UAS-Adk2-RNAi Adk2^GL00036 (FBti0144142), UAS-Ak1-RNAi Ak1^GL00177(FBti0144266), and control lines for RNAi y¹ v¹; P{CaryP}attP2 (FBst0036303), and y¹ v¹; P{CaryP}Msp300attP40 (FBst0036304) were obtained from the Bloomington Drosophila Stock Center. The SrpD-Gal4 strain (FBtp0020112) was obtained from M. Crozatier, backcrossed into the w¹¹¹⁸background, and recombined with P{tubP-GAL80ts}2 (FBti0027797), which was also backcrossed into w¹¹¹⁸ background, to generate the w¹¹¹⁸; +/+; SrpD-Gal4 P{tubP-GAL80ts}2 line with Gal4 expression in all hemocytes but not in the fat body when kept at 18°C for the first 3 days of development and then transferred to 25°C (S Fig; expression in the fat body is only present at 29°C in this line). All flies were grown on cornmeal medium (8% cornmeal, 5% glucose, 4% yeast, 1% agar, 0.16% methylparaben) at 25°C. Parasitoid wasps Leptopilina boulardi were reared on sugar agar medium (6% sucrose, 1.5% agar, 0.75% methylparaben) and grown by infection of wild-type Drosophila larvae. Early third instar larvae (72 hours after egg laying) were infected with parasitoid wasps (= time point 0 hours). Weak infection (1-2 eggs per larva) was used for survival analysis. Strong infection (4-8 eggs per larva) was used for the rest of the experiments to obtain a strong and more uniform immune response. Infections were performed on 60-mm Petri dishes with standard cornmeal medium for 15 minutes for weak infection and 45 minutes for strong infection.

Hemocyte counting

Drosophila hemocytes were counted at 20 hpi by bleeding one larva in 15 µL of PBS and using Neubauer hemocytometer (Brand GMBH) using differential interference contrast microscopy. Particular types of hemocytes were determined based on the cell morphology.

Survival and developmental time analysis

Infected and non-infected 3^rd instar larvae were placed into fresh vials in a number of 30 larvae per vial, 3 vials per genotype with 3 to 5 independent experiments. Differences in larval development were determined by counting newly appeared pupae every 6 hours post infection. Percentage of pupae number per total used larvae number at particular time points was plotted and compared by Log-rank test. For survival, all emerged adult flies were counted and checked if they contain the melanized eggs to confirm they were infected. The ones without any egg were excluded from the analysis. Percentage of emerged adults number per total used infected individuals were plotted.

Gene expression analysis

Based on the larval stage, 30 to 50 larvae were washed 3x in PBS, transferred on microscope slide covered with parafilm and placed on ice. All larvae were opened with tweezers and released hemolymph with hemocytes were aspired with 10 µL pipette and transferred into 1.5 mL tube containing TRIzol Reagent (Life Technologies) for RNA extraction. RNA was purified with Direct-zol RNA Microprep (Zymo-Research). Total RNA was used for cDNA preparation by PrimeScript Revrese Transcriptase (Takara) with oligo(dT) primer. cDNA was used as a template for qPCR using TP SYBR 2x Master Mix (Top-Bio) and C1000 Touch Thermal Cycler (Bio-Rad). Expression of a specific gene in each sample was normalized to expression of RpL32 (FBgn0002626) and relative gene expression was counted with ΔΔCt method. Graphs showing target gene expression and statistical analysis were made with GraphPad Prism.

Primers:

Bulk RNAseq analysis

RNA was extracted from circulating hemocytes (72 hours after egg laying = time of infection = 0 hours, 81 hours after egg laying = 9 hours post infection/hpi and 90 hours after egg laying = 18 hpi) of uninfected and infected 3^rd instar w¹¹¹⁸larvae and processed as described in details with all data in (Kazek et al. 2024). Raw data are available at The European Nucleotide Archive under study accession number: PRJEB74490 (secondary acc: ERP159178) (https://www.ebi.ac.uk/ena/browser/view/PRJEB74490).

Metabolomics and stable 13C isotope tracing in hemocytes ex vivo

Samples for metabolomics were collected 20 hours post infection. Larvae were washed first with distilled water and then with PBS to reduce contamination. Larval hemolymph was collected by carefully tearing the larvae on a glass microscope slide covered with parafilm. Hemolymph from 100 larvae was immediately collected into sterile 1.5 ml Eppendorf polypropylene centrifuge tubes prefilled with 100 µl PBS and centrifuged for 5 min at 25°C, 360xg. For experiments without 13C isotope tracing, hemocytes were resuspended in 50 µL of HPLC water and three times frozen in liquid nitrogen/thawed at 37 °C (established for sufficient cell disruption and extraction of intracellular metabolites). 200 µl of cold acetonitrile-methanol (1:1) was added and stored at −80 °C until LC-HRMS analysis described in Kazek et al. 2024. For experiments with 13C isotope tracing, the supernatant was removed and the cells were mixed with medium containing PBS, 5 mM trehalose, 0.5 mM glucose, 5 mM proline, 5 mM glutamine (Sigma/MERCK), supplemented with gentamicin (10 mg/ml; Gibco), amphotericin B (250 µg/ml; Gibco) and 0.1 mM phenylthiourea (PTU; Sigma/MERCK) to prevent melanization. Based on the experiment, medium further contained either 0,33 µM L-methionine-¹³C₅ or 10 µM adenosine-¹³C₅ and unlabeled 0,33 µM L-methionine. The hemocytes were then incubated ex vivo for 20 min at 25°C and 80-90% humidity. The cells were then centrifuged for 5 min at 25°C, 360xg, the supernatant was removed, the cells were mixed with 50 µl of cold PBS and frozen in liquid nitrogen/thawed at 37°C three times (to disrupt the rigid hemocyte cells). Finally, 200 µl of cold acetonitrile-methanol (1:1) was added, in addition, the internal standard 4-fluorophenylalanine (50 ng/sample) was added to control the extraction efficiency and mass spectrometer responses and samples were stored at −80°C until LC-HRMS analysis which is described in Kazek et al. 2024. When extracellular adenosine and inosine were measured, samples were collected in room temperature and centrifuged in 25°C, 5 min, 360xg to wash hemocytes from hemolymph and extracellular metabolites. Cells were resuspended in medium composed of PBS, 5 mM trehalose, 0.5 mM glucose, 5 mM proline, 5 mM glutamine, 0.33 mM methionine (Sigma/MERCK), supplemented with gentamicin (10 mg/ml; Gibco), amphotericin B (250 µg/ml; Gibco), 0.15 mM EHNA (Sigma/MERCK) and 0.1 mM phenylthiourea (PTU; Sigma/MERCK) to prevent melanization. Samples were incubated for 20 minutes in room temperature, and then centrifuged in 4°C, 5 min, 360xg. 70 µL of supernatant were transferred into new sterile 1.5 mL Eppendorf polypropylene centrifuge tube containing 70 µL of acetonitrile and stored at −80 °C until LC-HRMS analysis. Pelleted cells were stored in −80°C for protein measurement with QuantiPro^™ BCA Assay Kit (Sigma/MERCK).

Raw data are available at https://doi.org/10.6084/m9.figshare.27291300.v1 and processed data in S2_Data. For peak area analysis, the data were normalized to the total content of all screened unlabeled metabolites - the peak area of the metabolite in a particular sample was divided by the peak area of the same metabolite of the selected reference sample and this procedure was repeated for each individual unlabeled metabolite. These ratios of all metabolites in one particular sample were averaged to determine a normalization factor. We then divided the measured peak area by the normalization factor for that sample to obtain the normalized peak area values (normalization factors are reported in S2_Data).

Data analysis

Data were analyzed and graphed using GraphPad Prism (GraphPad Software), with specific statistical tests shown in the legend of each figure.

Acknowledgements

The authors acknowledge funding from the Grant Agency of the Czech Republic to TD (Projects 17-16406S and 20-09103S; www.gacr.cz) and from the Grant Agency of the University of South Bohemia in České Budějovice to PN (087/2019/P). We thank Dr. Michele Crozatier and Bloomington Drosophila Stock Center for fly and wasp stocks. We thank Dr. Michalina Kazek for help with metabolomics optimization, to Lucie Hrádková for laboratory management and Marcela Jungwirthová for project management, and all members of Doležal and Moos laboratories for their help with work. We thank Dr. Adam Bajgar for help with bulk RNAseq and Dr. Vladimír Beneš and Genomics Core Facility (EMBL Heidelberg, Germany) for RNAseq services and Dr. Bajgar and Dr. Gabriela Krejčová for help with optimization of the molecular biology methods and wasp infections and general manipulation with flies and larvae. We thank Dr. Ellen McMullen for manuscript editing.