We here report, for the first time, that rat pancreatic islet cells and the rat INS-1 β-cell line express detectable levels of transcripts encoding the key enzymes of the TKP. Importantly, this is associated with the spontaneous release of significant levels of the TKP metabolites KYN and KYNA by the rat islets, therefore suggesting some degree of functionality of the TKP in the rat islet cells in the basal situation. The basal expression levels across genes are maintained during in vitro cultureand likely reflect a stable and genuine genomic pattern.
Our results first show that IDO1 transcript and protein are not detectable in the rat pancreatic rat islet under basal unstimulated condition (as well as in the INS-1 β-cell line), similarly to the human islet . In our study, we demonstrated that IFN-ï§ï€ alone strongly activates IDO1 mRNA accumulation in rat islets. IL-1β also induced a robust increase of IDO mRNA accumulation. By contrast, exposure to HGHF or H2O2 did not activate IDO expression, at least under our experimental setting which was nevertheless sufficient to deteriorate INS-1 cell survival and glucose-stimulated insulin secretion. effect of IFN-ï§ï€ on IDO1 transcripts have been documented also in human islets by microarray profiling experiments [33,39] and the pattern of IFN-ï§ï€ activated IDO expression encompasses a variety of cell types, including β and nonβ-cells . Concerning IDO1 protein expression, we must mention that our FACS-purified cell experiments suggest that IFN-ï§ï€ activates accumulation of IDO protein in both rat β-cells and nonβ-cells, and it is confirmed by an increased production of KYN in treated islets. But the specific immunostaining signal remained faint. This may reflect a very low expression of the IDO1 protein, which can be confirmed by a lack of significant change in the Trp consumption in IFN-ï§, treated islets. Although there is a 1.4 folds increase of the KYN production in IFN-ï§ treated isletsï€ , the increased amount corresponds only 3% of consumed Trp by islets. More convincing to support the involvement of IDO1 activation following treatment with IFN-ï§ï€¬ï€ is our finding that IFN-ï§ï€ induced anï€ increased KYN release by the rat islets while KYNA release remained unchanged.
Concerning IL-1β, a recent report demonstrated that it induced an up-regulation of IDO1 transcripts in human hippocampal progenitor cells , an observation consistent with our finding. In addition, our results show a faint IDO1 induction by corticosterone, as we detected a very tiny amount IDO1 amplicon in the corticosterone- treated samples compared to zero amplicon in the untreated group. However, due to poor reproducibility, the result did not reach statistical significance.
TDO2 transcripts and protein were moderately expressed in the rat pancreatic rat islet under basal unstimulated condition (as well as in the INS-1 β-cell line). IFN-ï§, IL-1β, HGHF, or the combination IFN-ï§ï€ + HGHF, induced a decrease of the level of TDO mRNA in rat islets, which became significant in response to IL-1β. Expression of TDO2 was also reduced by IFN-ï§ in neuroblastoma cells . Exposure to corticosterone did not activate islet TDO2 expression (either transcripts or protein), at variance with the effect reported in liver. Lack of a suitable antibody did not allow immunohistochemical studies of TDO2 protein.
KYN, The first stable metabolite, can be switched into two different routes: either catalyzed by the KATs to form KYNA (KYNA branch), or by KMO, which may lead to the production of QUIN and NAD (QUIN/NAD branch) .
In the KYNA branch of the TKP, four subtypes of KAT have been identified within the brain, with KAT II being the most abundant in rat and human brain . The four KAT enzymes were found constitutively expressed in rat pancreatic islets as well as in the INS-1 cells, with KAT transcripts detected at the highest level compared to the other TKP transcripts. Except a faint increase of KAT3 transcript, we observed no effect, or only marginally either increased or decreased levels of transcripts encoding the other KAT enzymes in response to IFN-ï§ï€ exposure. Our observations are in line with the effects of IFN-ï§ï€ observed in human fibroblasts  or neuronal cells , In neuroblastoma cells, the IFN-ï§ï€ treatment did not modify the levels of transcripts encoding KAT1 and KAT2 .
In the NAD branch of the TKP, KMO convert KYN to 3-hydroxykynurenine (3HK), which is further degraded to 3-hydroxyanthranilic acid (3HAA) through the action of Kase. After HAAO, the metabolism in the liver enters either into the complete oxidation pathway through which ATP is formed, or into QUIN which is finally degraded into nicotinamide adenine dinucleotide (NAD). From the complete oxidation pathway, a small amount of aminomuconic semialedhyde (AMS) is also formed through the action of aminocarboxymuconate semialedhyde decarboxylase (ACMSD). In physiological condition, the TKP mainly results in ATP formation and only small amount of NAD is formed through the action of quinolinic acid phosphoribosyltransferase (QPRT) .
We found that rat pancreatic islets (but not INS-1 cells) constitutively express KMO under basal unstimulated condition. 48h-exposure of rat islets to HGHF induced an upregulation of transcripts for KMO (7.6 fold; p<0.05). IL-1β also upregulated the KMO transcripts (19 fold; p<0.01) and KMO protein (2.5 fold; p<0.05). By contrast, IFN-ï§ï€ exposure (or H2O2) did not significantly affect KMO mRNA, under our experimental setting, and it did not potentiate the HGHF-induced KMO expression. Our immunohistochemical data suggest that HGHF activated accumulation of KMO protein mostly in rat β-cells as shown by our FACS-purified cell experiments. This is in accordance with the data related to KMO mRNA accumulation in basal or IFN-ï§+HGHF conditions, in the β-cell fraction. In human skin fibroblast cultures, a lack of effect on transcripts encoding KMO following IFN-ï§ï€ treatment  was reported, but a small increase in KMO activity in IFN-ï§ï€ treated monocytes was also reported . Concerning IL-1β, it induced an up-regulation of KMO transcripts in human hippocampal progenitor cells . Moreover, as we have detected an increased production of IL-1β by HGHF treated islets themselves (data not show). The effect of the induction of KMO by HGHF is possible though this auto-secretion of IL-1β.
Rat pancreatic islets (but not INS-1 cells) constitutively express Kase under basal unstimulated condition. FACS-purified cell experiments indicated that Kase transcripts were present in β-cells as in nonβ- islet cells. Levels of transcripts encoding Kase enzyme were not significantly modified in response to IFN-ï§, IL-1β, HGHF, H2O2 or the combination HG/HF + IFN-ï§.
A similar pattern of expression was observed for QPRT, with detectable QPRT transcripts in β-cells as in nonβ-cells in basal condition, no significant alteration in response to IFN-ï§, IL-1β, H2O2 or the combination HGHF + IFN-ï§, and a moderate but significant increase in response to HGHF.
Rat pancreatic islets and INS-1 cells did not express ACMSD either in basal unstimulated condition or in response to IFN-ï§, IL-1β, HGHF, H2O2, or the combination HGHF + IFN-ï§. This finding suggests that the islets and INS1 cells are incapable to completely degrade Trp for the ATP synthesis. The QUIN branch of TKP leads only the formation of NAD +.
Rat pancreatic islets as well as INS-1 cells constitutively express NAMPT under basal unstimulated condition. FACS-purified cell experiments indicated that NAMPT transcripts were present in β-cells as in nonβ-cells. Levels of transcripts encoding NAMPT enzyme were upregulated in response to IFN-ï§ï€ ï€¨1.6 fold pï€¼°ï€®°ï€±ï€©, HGHF (2.4 fold; pï€¼°ï€®°ï€µï€©, IL-1β (2.3 fold; pï€¼°ï€®°ï€±), while they remained unchanged after H2O2. At variance with lack of data related to expression and effect of TKP enzymes upstream NAMPT in islet cells, there are several reports indicating that NAMPT expression is detected in human islets (both in β- and nonβ- islet cells) and increased by high glucose [45,46]. This is in line with our present findings in rat islets. NAMPT exists in two known forms : an intracellular NAMPT (iNAMPT) and a secreted form, extracellular NAMPT (e NAMPT). iNAMPT is a rate-limiting enzyme in the mammalian NAD+ salvage. But eNAMPT is an hormone-like molecule, known as both pre-B cell colony enhancing factor (PBEF) due to its cytokine-like function ,and visfatin due to its adipokine function. eNAMPT may also convert extracellular nicotinamide to nicotinamide mononucleotide (NMN). NMN functions as an extracellular metabolite that may regulate the insulin secretion in β-cell. [Revollo et al. Nampt/PBEF/Visfatin Regulates Insulin Secretion in β Cells as a Systemic NAD Biosynthetic Enzyme]. .
Finally, to summarize our present findings, while the enzyme IDO1 which is the first limiting step in the catabolism of tryptophan, is not expressed, the normal rat islet constitutively expresses transcripts encoding the different enzymes of the KYNA branch of the TKP (KATs). It also express enzymes of the QUIN/NAD branch (KMO, Kase, QPRT, NAMPT), while the lack of ACMSD expression makes PIC or glutaryl-CoA production unfeasible. Our data also show that islet IDO1 and KMO are main targets regulated by environmental conditions associated to obesity /diabetes, since IDO1 expression is activated by cytokines (but not by HGHF or oxydative stress) while KMO expression is prominently activated by HGHF and IL-1ï¢ï€ (but not or to lesser extent, by IFN-ï§, oxydative stress or glucocorticoid). The many discrepancies identified here between islet cells and other cell types in response to cytokines or HGHF, reflect probably intrinsic differences across cell types in transcription of genes encoding TKP enzymes and in their abilities to form active TKP metabolites such as KYN, KYNA or QUIN . The enzymes participating in the TKP, when expressed, are distributed at similar levels in both rat β-islet cells and nonβ-islet cells, except alpha cells that express more KMO and QPRT. The specialization of TKP in different types of cells is not a unique phenomenon in the islets. Also in the central nervous system, the TKP enzymes are found differently distributed among the different cells: the astrocytes express little KMO while the microglial cells are lack of KAT . Astrocytes are therefore primarily responsible for the synthesis of KYNA, while production of QUIN is predominantly attributed to microglial cells .
The physiological role of the TKP in pancreatic islets is not known. Since the first description of the TKP in CNS thirty-some years ago, this pathway has been at the center of stage in virtually all major CNS diseases. Indeed, it may serve as a unique interface between the neuronal signaling and immune system , since it release some neuromodulatory metabolites in response to inflammatory signaling. Moreover, recent studies in yeast, worm, flies and mice have now established that TKP metabolism can also function as a regulator in lifespan and cancer . In addition to its participation in NAD+ homeostasis, the TKP generates a number of biologically active intermediate metabolites. QUIN is an excitotoxic agent at glutamate receptors and synergizes with 3-HK to produce ROS. KYNA is an antagonist of glutamate neurotransmission and nicotinic α7 receptors. In addition to NMDA, AMPA and nicotinic a7 receptors, KYNA is reported to interact with GPR35  and aryl hydrocarbon receptors , but its physiological roles are still not clear.
Support for a tolerogenic role for TKP activation in islets, the adenovirus-delivered IDO1 overexpression in transplanted NOD islets improves their survival in NOD/SCID recipients that received diabetogenic splenocytes [48,49]. It was postulated that short-term IDO1 activation might protect islets, on one hand, by reducing cytotoxic damage through its superoxides consuming activity and perhaps reducing MHC class 1 expression in islet. On the other hand, by perturbing of the function of activated T and B-lymphocytes through local depletion of free Trp. In addition KYN, KYNA and PIC may also directly inhibit T-cell proliferation [48,49]. The release of Trp metabolites such as 3-HK, 3-HAA, QUIN and PIC could also provide bystander inhibition of immune cell function.
Moreover, since pancreatic β-cells express functional glutamate receptors , an effect of β-cell exposure to QUIN or KYNA may be relevant in this regard. In fact, the effects of β-cell exposure to QUIN or KYNA are so far unknown. 3-HK and 3-HAA have been reported to acutely inhibie leucine-stimulated insulin secretion from rat islets . The effects of -cell exposure to KYN have not been reported.
In conclusion, we have demonstrated for the first time in normal rat pancreatic islets, that: 1/only some TKP genes are constitutively expressed, both in β-cells as well as nonβ- cells; 2/ the rate-limiting enzyme IDO1 is not constitutively expressed; 3/ IDO1 and KMO expression are potently activated by proinflammatory cytokines (IFN-ï§, IL-1ï¢) or glucolipotoxicity respectively, rather in β-cells than in nonβ-cells; 4/ oxidative stress or glucocorticoid modulate TKP genes only marginally. Therefore the pancreatic islets may represent a new target tissue for inflammation and glucolipotoxicity to activate the TKP pathway. Of course the next step will be to evaluate the functional impact the various TKP metabolites have upon the islet cells. Most importantly, since inflammation is now recognized as a crucial mechanism in the development of the metabolic syndrome  and more specifically at the islet level , it is therefore needed to evaluate the potential induction of the TKP in the endocrine pancreas during obesity and/or diabetes and its relationship to the islet cell functional alterations.