*3.4. HPP Treatment Inhibits the Maturation of LPS-Stimulated Human DC, Resulting in Reduced Activation of CD4+ T Cells*

In order to determine if the ketoacids impact DC maturation (and in turn T cell activation), human DC were treated with HPP (500–1000 μM) for 6 h prior to stimulation with LPS (100 ng/mL) for 24 h. Surface expression of maturation and co-stimulatory markers (CD40, CD80, CD83 and CD86) were measured by flow cytometry (due to the fluorescent nature of IP it is unsuitable for this flow cytometric analysis and hence was not included in these experiments). As expected, the expression of co-stimulatory markers was increased in LPS-stimulated DC. The average Median Fluorescent Intensity (MFI) of these cells was set to 100% and used as a control for comparison to the HPP-treated cells. There was a significant decrease in MFI at both HPP concentrations when compared to the LPS-stimulated control (Figure 4A). The phagocytic capacity of DC was next measured upon incubation of DC with FITC-conjugated DQ-Ovalbumin (DQ-Ova; 500 ng/mL) and uptake of the model antigen was assessed by flow cytometry. Compared to untreated cells, the ability of LPS-treated DC to phagocytose DQ-OVA was significantly impaired, signifying a heightened maturation status, however, pre-incubation with HPP prior to LPS treatment attenuated this effect and maintained the DC in an immature state (Figure 4B). Finally, and in order to determine if the reduced DC activation status that occurs in the presence of HPP has an impact on T cell activation, DC were treated with HPP (1000 μM) for 6 h prior to stimulation with LPS (100 ng/mL) for 24 h. The cells were then incubated with CD4+ T cells at a ratio of 10:1 (CD4+ T cells:DC) for five days. The supernatants were removed for cytokine analysis by ELISA and cells were stimulated with PMA (50 ng/mL), ionomycin (500 ng/mL) and brefeldin A (5 μg/mL) for 4 h. Expression of ki67, which is a measure of cell proliferation, was assessed by flow cytometry in CD3+CD8<sup>−</sup> cells (gating strategy shown in Supplementary Figure S4), as was production of the pathogenic Th1 cytokine, IFNγ. T cells co-cultured with HPP-treated, LPS-stimulated DC showed a trend towards reduced ki67 and IFNγ expression when compared to T cells co-cultured with LPS-stimulated DC alone (Figure 4C). Interestingly, the T cells exhibited a trend towards enhanced production of the anti-inflammatory cytokine IL-10, when compared to T cells co-cultured with LPS-stimulated DC alone (Figure 4D). Overall, these results indicate that HPP reduces the maturation of innate DC, which in turn impacts adaptive T cell responses and may skew them towards a more anti-inflammatory phenotype.

**Figure 3.** HPP and IP reduce the production of pro-inflammatory cytokines in LPS-stimulated human DC. Primary human DC were left untreated (UT) or incubated with IP (**A**,**C**,**E**,**G**,**I**) or HPP (**B**,**D**,**F**,**H**,**J**) (500–1000 μM) for 6 h prior to stimulation with LPS (100 ng/mL) for 24 h. Cell supernatants were assessed for TNF, IL-6, IL-23, IL-12p70, and IL-10 secretion by ELISA. Pooled data depict mean (±SEM) cytokine concentrations for four to seven healthy donors (means of three technical replicates per donor). Repeated measures one-way ANOVA, with Dunnett's multiple comparisons post hoc test, was used to determine statistical significance, by comparing means of treatment groups against the mean of the control group (\*\* *p* < 0.01, \* *p* < 0.05).

**Figure 4.** HPP treatment reduces DC maturation and subsequent CD4<sup>+</sup> T cell activation. Primary human DC were left untreated (UT) or incubated with HPP (500–1000 μM) for 6 h prior to stimulation with LPS (100 ng/mL) for 24 h. (**A**) Cells were stained for CD40, CD80, CD86, and CD83 and analysed by flow cytometry. Histograms showing the expression of maturation markers for HPP-treated, LPS-stimulated DC compared to unstimulated cells or LPS stimulation alone from one representative experiment. Pooled data showing the mean (±SEM) MFI for each marker expressed as a percentage of control (LPS stimulation alone) from six to seven healthy donors. (**B**) DC were incubated with FITC-conjugated DQ-Ovalbumin (DQ-Ova; 500 ng/mL) for 20 min and were immediately acquired by flow cytometry. Dot plots depicting DQ-Ova uptake from one representative experiment. Pooled data showing the mean (±SEM) DQ-Ova uptake as a percentage of total cells from nine healthy donors. (**C**) DC were pre-treated with HPP prior to stimulation with LPS, and subsequently cultured with CD4+ T cells for five days. Dot plots depicting ki67 expression (as a measure of proliferation) and IFNγ expression from one representative experiment. Pooled data showing the mean (±SEM) of ki67+ and IFNγ<sup>+</sup> cells as a percentage of CD3+CD8<sup>−</sup> cells from four healthy donors. (**D**) Cell supernatants were assessed for IL-10 secretion by ELISA. Pooled data depict mean (±SEM) cytokine concentrations for four healthy donors (means of three technical replicates per donor). Repeated measures one-way ANOVA, with Dunnett's multiple comparisons post hoc test, was used to determine statistical significance by comparing means of treatment groups against the mean of the control group (\*\*\* *p* < 0.001, \*\* *p* < 0.01, \* *p* < 0.05).

#### *3.5. HPP and IP Modulate Metabolic Reprogramming in LPS-Stimulated Human DC*

Metabolic reprogramming has been observed in immune cells and numerous recent studies have demonstrated that their activation/maturation is accompanied by a metabolic switch favouring glycolysis over oxidative phosphorylation [21]. In order to determine if *T. brucei*–derived ketoacids have an effect on immune cell metabolism, DC were pre-treated with IP or HPP (1000 μM) for 6 h prior to stimulation with LPS (100 ng/mL) for 12 h. These cells were then analysed in a Seahorse XFe96 analyser following the addition of oligomycin (1 mM), an inhibitor of mitochondrial complex V, FCCP (1 mM), a mitochondrial uncoupler, rotenone (500 nM) and antimycin A (500 nM), which are inhibitors of the mitochondrial complexes I & III, respectively, and 2-DG (25 mM), an inhibitor of glycolysis. The metabolic activity of the cells was then determined by measuring the ECAR, which is a measure of glycolysis, and the OCR, which is a measure of oxidative phosphorylation. IP- and HPP-treated DC showed no change in basal glycolysis when compared to LPS stimulation alone (Figure 5B). LPS-treated cells showed a trend towards increased max glycolysis, and this was significantly decreased in the presence of either HPP or IP (Figure 5C). Both IP and

HPP were also capable of significantly decreasing the glycolytic reserve in LPS-stimulated cells (Figure 5C). There were no significant changes in the basal respiration (Figure 5F), max respiration (Figure 5G), and respiratory reserve (Figure 5H) in IP- or HPP-treated DC when compared to LPS stimulation alone, suggesting that they have no impact on oxidative phosphorylation.

**Figure 5.** HPP and IP modulate metabolic reprogramming in LPS-stimulated DC. Primary human DC were pre-treated with either HPP or IP at 1000 μM for 6 h before stimulation with LPS (100 ng/mL) for 12 h. The extracellular acidification rate (ECAR) and the oxygen consumption rate (OCR) were measured using a Seahorse XFe96 analyser before and after the injections of oligomycin (1 mM), FCCP (1 mM), antimycin A (500 nM) and rotenone (500 nM), and 2-DG (25 mM). Bioenergetic profiles from one representative experiment depicting (**A**) ECAR and (**E**) OCR measurements over time. Pooled data (*N* = 6) depicts the calculated mean (±SEM) of (**B**) basal glycolytic rate, (**C**) max glycolytic rate, (**D**) glycolytic reserve, (**F**) basal respiratory rate, (**G**) max respiratory rate, and (**H**) respiratory reserve for each treatment group. (**I**) HK2 expression was measured by western blot. Densitometry results shown are mean ± SEM of the relative expression of HK2: β-actin from five to seven healthy donors. (**J**) FLIM images of DC measuring intracellular NADH. Pooled data (*N* = 4) depicts the mean (±SEM) of the ratio of bound:free NADH, represented by the *τ* average. Repeated measures one-way ANOVA, with Dunnett's multiple comparisons post hoc test, was used to determine statistical significance by comparing means of treatment groups against the mean of the control group (\*\* *p* < 0.01, \* *p* < 0.05). ImageLab (Bio-Rad) software was used to perform densitometric analysis.

Expression of HK2, the rate limiting enzyme in the glycolytic pathway, was next assessed by western blotting. HK2 is known to be induced by inflammatory stimuli and, as expected, LPS stimulation induced the upregulation of HK2 in DC. However, there was reduced expression of the enzyme in IP-treated DC, and a trend towards reduced expression (albeit not significant) in HPP-treated DC when compared to LPS-stimulated DC (Figure 5I). The effects of HPP treatment on the metabolism of DC was further investigated using FLIM, which measures the intracellular levels of NADH. Bound NADH, which is associated with oxidative phosphorylation, or free NADH, which is associated with glycolysis, can be distinguished based on their distinct lifetimes upon fluorescence excitation. The ratio of

bound to free NADH can be used to measure whether a cell is favouring the engagement of glycolysis (a decrease in the ratio due to increased free NADH) or oxidative phosphorylation (an increase in the ratio due to increased bound NADH). Similar to the Seahorse results reported above, LPS-stimulated DC ramped up glycolysis, as represented by a decrease in the *τ* average compared to untreated DC (Figure 5J). Cells pre-treated with HPP exhibited a significant increase in the *τ* average compared to the LPS-stimulated controls, indicating they are favouring oxidative phosphorylation to generate their energy (Figure 5J). Overall, these results indicate that *T. brucei*-derived ketoacids modulate DC metabolism, reducing engagement of glycolysis, which is associated with rapid inflammatory responses.

#### *3.6. HPP and IP Activate Autophagy-Related Proteins*

We have previously demonstrated that the HO-1 inducers carnosol and curcumin activate the autophagy regulator AMPK, which incidentally is also known to downmodulate glycolysis in immune activated cells [22]. Furthermore, the key autophagy-related protein, p62, is linked to Nrf2 activation [23–25]. In order to determine if *T. brucei*-derived ketoacids have any impact on autophagy-related proteins, DC were treated with IP or HPP (both 1000 μM) for 15 min and phosphorylation (and therefore activation) of AMPK was assessed by western blotting. IP treatment resulted in a significant increase in AMPK phosphorylation while there was a trend towards increased activation with HPP (Figure 6A). DC treated with IP also showed an increase in both p62 and LC3-II (which is converted from LC3-I during autophagy) over time, and this was most potent after 24 h (Figure 6B,C). HPP treatment significantly increased p62 expression after 6 h and significantly increased LC3-II expression after 24 h (Figure 6B,C). These results indicate both ketoacids are activating autophagy-related proteins in human DC.

#### *3.7. HPP and IP Reduce Proliferation and Cytokine Expression in PBMC Isolated from IBD Patients*

It has previously been reported that IP has powerful immune suppressive effects in a murine experimental colitis model [7]. In order to determine if these results translate to a more clinical setting, PBMC were isolated from IBD patients and treated with IP or HPP (250–1000 μM) for 6 h prior to stimulation with anti-CD3 (1 μg/mL) for a further four days. Furthermore, the culture media was replaced with fresh media after 18 h of incubation with the compounds to circumvent issues surrounding the fluorescence of IP over long periods of time. The supernatants were removed for cytokine analysis by ELISA and cells were stimulated with PMA (50 ng/mL), ionomycin (500 ng/mL), and brefeldin A (5 μg/mL) for 4 h. Expression of ki67, and the cytokines IFNγ and IL-17 (both of which are known to play a pathogenic role in IBD), were measured by flow cytometry in CD3+CD8<sup>−</sup> cells (gating strategy shown in Supplementary Figure S5). IP was non-toxic to PBMC at all concentrations tested, however there was some toxicity seen when using the higher concentrations of HPP (Supplementary Figure S6). Despite this being significant, it is unlikely to account for the effects seen, as flow markers were examined in live cells only. Both IP- and HPP-treated cells were capable of dose-dependently reducing the proliferation and expression of IFNγ when compared to anti-CD3 stimulation alone (Figure 7A,B), while having no effect on the intracellular levels of IL-17. However, when cell supernatants were assessed by ELISA, there was a significant reduction in both IFNγ and IL-17 production in IP- and HPP-treated cells (Figure 7C,D). Unlike purified DC, the ketoacids did not enhance the production of IL-10 from PBMC, suggesting a cell-type specific effect (Figure 7C,D).

**Figure 6.** HPP and IP modulate autophagy-related proteins. (**A**) Primary human DC were left untreated (UT) or incubated with IP or HPP at 1000 μM for 15 min. Phosphorylation of AMPK was measured by western blot. Densitometry results shown are mean ± SEM of the relative expression of p-AMPK: β-actin from four healthy donors. (**B**,**C**) Primary human DC were left UT or incubated with IP or HPP at 1000 μM for 6, 12, or 24 h. Expression of (**B**) p62 and (**C**) LC3 were measured by western blot. Densitometry results shown are mean ± SEM of the relative expression of (**B**) p62: β-actin from five healthy donors and (**C**) LC3 II: β-actin from six healthy donors. (**A**) Statistical significance was determined using a Paired *t*-test. (**B**,**C**) Statistical significance was determined by repeated measures one-way ANOVA with Dunnett's multiple comparisons post hoc test to compare means of treatment groups to the control group (\*\*\* *p* < 0.001, \*\* *p* < 0.01, \* *p* < 0.05). ImageLab (Bio-Rad) software was used to perform densitometric analysis.

**Figure 7.** HPP and IP reduce proliferation and cytokine expression in ex vivo stimulated PBMC from patients with Inflammatory Bowel Disease. PBMC isolated from IBD patients were treated with (**A**,**C**) HPP or (**B**,**D**) IP (250 μM–1000 μM) for 6 h prior to stimulation with anti-CD3 for 12 h. After 18 h, culture media was replaced with fresh media and cells were incubated for a further 4 days with anti-CD3 stimulation. Supernatants were removed for analysis of cytokine concentration by ELISA. (**A**,**B**) Proliferation and cytokine production by CD3+CD8<sup>−</sup> cells was analysed by flow cytometry. Pooled data (*N* = 14) depicting the mean ± SEM of ki67 (as a measure of proliferation), IFNγ, and IL-17 in CD3+CD8<sup>−</sup> T cells. (**C**,**D**) Cell supernatants were assessed for concentrations of IL-10, IFNγ, and IL-17 by ELISA. Pooled data depicts mean (±SEM) cytokine concentrations for six IBD patients (means of three technical replicates per donor). Statistical significance was determined by repeated measures one-way ANOVA with Dunnett's multiple comparisons post hoc test to compare means of treatment groups to the control group (\*\*\* *p* < 0.001, \*\* *p* < 0.01, \* *p* < 0.05).

#### **4. Discussion**

The production of large amounts of immunomodulatory aromatic ketoacids during *T. brucei* infection likely serves to benefit the parasite by prolonging infection, proliferation, and, ultimately, survival in the host. While the secretome of *T. brucei* has been shown to reduce the secretion of IL-12, IL-10, IL-6, and TNF in both murine and human DC [26,27], the ketoacids, IP and HPP, have been shown to directly ameliorate inflammatory cytokine production in murine macrophages and glia [4,5]. This is further supported by studies demonstrating their therapeutic efficacy in murine models of disease [4,6,7]. Given the key role played by DC in shaping both innate and adaptive immune cell responses, we sought to determine if these effects translate to this vital human immune cell population, which not only serves to present antigens during infection, but also plays a key role in determining pathogenic T cell responses during disease. We demonstrate that both IP and HPP are capable of significantly reducing the secretion of a number of pro-inflammatory cytokines, including TNF, IL-6, IL-12, and IL-23 in LPS-stimulated DC. Furthermore, the ketoacids upregulate HO-1 in an Nrf2 dependant manner, which is in line with studies demonstrating that HO-1 induction can promote a more tolerogenic DC phenotype [18,19,28]. In support of this, we also demonstrate that ketoacid-treated DC have a reduced capacity to activate T cells, which in turn limits the production of the pathogenic T cell cytokine IFNγ, which is known to play a deleterious role in a number of inflammatory/autoimmune conditions including IBD [29].

In comparison to the previous data from murine macrophages and glia, a larger repertoire of cytokines is inhibited by the ketoacids in human DC. For example, IP and HPP had no effect on TNF secretion in murine glia and no effect on either TNF or IL-6 secretion in bone marrow derived macrophages [4,5], however, these cytokines were significantly reduced in LPS-activated human DC. Furthermore, IP directly induced the production of the anti-inflammatory cytokine IL-10 in DC. The maturation status of DC is also impacted by ketoacid treatment, with the HPP-treated cells exhibiting a reduction in co-stimulatory and maturation markers, which, in turn, prevents their ability to participate in T cell activation. The upregulation of HO-1 was also accompanied by the expression of additional Nrf2-regulated genes and both IP and HPP appeared to exhibit direct antioxidant activity, which may explain their ability to activate Nrf2, given its well documented capacity to rapidly respond to oxidative stress.

Further analysis of these novel HO-1 inducers also revealed that they can modulate immune cell metabolism. Indeed, recent studies have highlighted an important link between immune cell activation and metabolism. It is now well recognised that, not only do different immune cells engage different metabolic pathways, but that the activation/maturation state of the immune cells is accompanied by metabolic switches. For example, innate immune cells, including macrophages and DC, ramp up glycolysis in order to rapidly generate sufficient energy and the building blocks required to fight infection [21]. This phenomenon is also a feature of pathogenic immune cells, and a significant effort is underway to determine if controlling/preventing dysregulated metabolic reprogramming can serve to ameliorate detrimental immune cell activation during disease. Here we demonstrate that both IP and HPP can decrease the max glycolysis observed in LPS-treated DC while also downregulating the expression of hexokinase 2, the rate-limiting enzyme in glycolysis. These results are similar to our recent observations with the HO-1 inducers, carnosol and curcumin, suggesting that HO-1 may be an important regulator of immune cell metabolism [30]. Further study is required to determine if additional features of metabolism are affected by ketoacids and whether metabolic reprogramming occurs during *T. brucei* infection, but these in vitro results are in line with the notion that metabolism is intricately linked with immune cell activation, and that the downmodulation of glycolysis in immune cells promotes a more tolerogenic phenotype.

From a therapeutic stance, IP in particular has shown potential in a murine model of colitis where administration of the ketoacid not only improved disease outcome, but also decreased expression of pro-inflammatory cytokines, including IL-12, IFNγ, and TNF [7]. We observed similar results in PBMC from IBD patients where both HPP and IP reduced the proliferation of anti-CD3 stimulated T cells, as well as the secretion of both IL-17 and IFNγ. Indeed, induction of HO-1 is being explored as a therapy for IBD and has shown promise in a number of murine models of disease [31–35]. Patients are currently treated largely with anti-inflammatories including 5-aminosalicylic acid (5-ASA), corticosteroids, methotrexate, and anti-TNF therapies [36]. However, many patients are/become refractory to these treatments and will require surgery in their lifetime. While further in vivo study is required to fully elucidate their efficacy and provide further information regarding treatment route, dosing, and long-term effects, these findings provide further impetus to explore aromatic ketoacids as a treatment for IBD (and indeed other inflammatory diseases), either alone or in combination with existing therapies.

Finally, a particularly interesting finding of this study is the observation that IP and HPP can activate key autophagic proteins. Autophagy itself is carried out by a number of autophagy-related (Atg) proteins and initiation is under the control of the protein kinases mTOR and AMPK, which are both intrinsically linked to immune cell metabolism (AMPK inhibits glycolysis while mTOR activation is linked to induction of glycolysis). The autophagic process is complicated and involves many different proteins and has been reviewed in detail elsewhere [37–39]. Briefly, a complex of Atg proteins lipidates LC3-I converting it to LC3-II [37,38]. LC3-II binds to the autophagosome membrane and facilitates the docking of cargos and proteins for degradation through their binding to p62 [40]. Following maturation, the autophagosome fuses with the lysosome to form the autolysosome, the contents of which are then degraded and recycled. In this study, we show that IP and HPP activate AMPK and increase expression of p62 and LC3-II. Autophagy is of particular importance for DC, as many of their key functions, including antigen uptake and presentation, are strongly associated with autophagy [41]. Despite some conflicting reports, generally it appears that activation of autophagy gives rise to a more tolerogenic DC phenotype, exhibiting reduced antigen presentation and maturation [41], which is similar to the results seen in this study. Furthermore, AMPK activation has been shown to attenuate pro-inflammatory cytokine production and DC maturation [22,42]. Therefore, ketoacid-induced AMPK and subsequent autophagy activation may serve to downmodulate DC maturation, antigen presentation, and glycolysis.

Notably, p62 is not only an important autophagy-related protein, it also plays a crucial role in the activation of Nrf2 [23–25]. Nrf2 is activated upon release from KEAP1, which can occur when p62 sequesters KEAP1, targeting it for degradation and allowing Nrf2 to translocate to the nucleus [23–25]. Therefore, the activation of p62 by both ketoacids, but in particular IP, may be responsible for the subsequent activation of Nrf2, and therefore HO-1, by these ketoacids. While HPP induces HO-1 at an earlier time than IP, the latter appears to have more potent effects overall. In most cases, the in vitro effects of IP are most apparent at 24 h, and we cannot rule out the possibility that it is converting to another form over time. Further study is undoubtedly required to delineate the true impact of aromatic ketoacids (and their potential derivatives) and HO-1 induction on autophagy-related processes, in addition to the noted effects on DC metabolic reprogramming.

#### **5. Conclusions**

In conclusion, the data presented here expands our understanding of the mechanism of action of *T. brucei*-derived ketoacids in human immune cells and suggests that HO-1 induction may be useful to regulate the metabolism and, therefore, function of immune cells in inflammatory disease. We firmly believe that these compounds represent novel and exciting HO-1 inducers worthy of further exploration.

**Supplementary Materials:** The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/antiox11010164/s1, Figure S1: HPP and IP are non-toxic to human DC; Figure S2: IP and HPP are not contaminated with endotoxin; Figure S3: ML385 is non-toxic to human DC; Figure S4: Gating strategy used to generate data shown in Figure 4; Figure S5: Gating strategy used to generate data shown in Figure 6; Figure S6: IP is non-toxic to PBMC, while higher concentrations of HPP shows significant, albeit mild, reductions in viability; Table S1: Seahorse calculations.

**Author Contributions:** H.K.F., J.M.F. and A.D. conceptualized and designed experiments. H.K.F. performed all experiments with laboratory assistance and intellectual input from S.A.O. and E.D. N.G.B.N. performed the FLIM experiments. N.G.B.N. and M.G.M. contributed to the interpretation of the FLIM data. D.P.N. contributed to the interpretation of data. M.T., J.D., E.J.R. and G.A.D. provided access to IBD patient samples. H.K.F., J.M.F. and A.D. wrote, reviewed, and edited the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Health Research Board, Ireland, grant number ILP-POR-2017-041.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki, and approved by the Research Ethics Committee of the School of Biochemistry and Immunology, Trinity College, Dublin (BI-HF-300921, approved 18 October 2018).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data is contained within the article or supplementary material.

**Conflicts of Interest:** The authors declare no conflict of interest.
