**5. Participation of L-arginine/L-glutamine coupling in diabetes**

From the above discussion, it seems clear that the development of diabetes is not simply a question of cytokine imbalance culminating in a redox disruption and consequent oxidative stress that disrupts or kills -cells. This, in fact, raises another question: is beta cell susceptibility to stress solely a question of compromised antioxidant defence? If this were the case, it would appear preposterous that such a sophisticated cell remains prone to endogenously-generated NO-mediated self-destruction. The intricate metabolism of L-arginine in -cells may unravel some important points in this regard.

In -cells, pro-inflammatory cytokines induce the production of NO, synthesised from L-arginine, via a reaction catalysed by iNOS, whose functionality depends on NF-B-driven gene transcription and *de novo* enzyme synthesis. iNOS also utilises NADPH and O2 as co-

A Novel L-Arginine/L-Glutamine Coupling Hypothesis: Implications for Type 1 Diabetes 255

*N*G-hydroxy-L-arginine (L-NOHA), an intermediate in the biosynthesis of NO, is a potent competitive inhibitor of L-arginase I (Boucher et al., 1994; Daghigh et al., 1994). Indeed, substantial amounts of this metabolite are released by LPS-treated rat alveolar macrophages (Hecker et al., 1995), while inhibition of L-arginase by L-NOHA may ensure sufficient availability of L-arginine for high-output production of NO in activated cells. L-Citrulline, the co-product of iNOS catalysis, and *S-*nitrosoglutathione (SNOG), an adduct produced by the reaction of NO with GSH, are also inhibitors of L-arginase in many cell types (Daghigh et al., 1994; Knowles & Moncada, 1994), including -cells (Cunningham et al., 1997). Hence, intermediates of NO synthesis, as well as NO itself, precisely coordinate a maximum of flux through iNOS in insulin-producing pancreatic cells (**Fig. 1**). Conversely, dexamethasone and dibutyryl cAMP block both iNOS and L-arginase expression, which is paralleled by a strong decrease of NO production (Gotoh & Mori, 1999). Additionally, macrophages treated with LPS and IFN undergo NO-dependent apoptosis, which may be prevented by L-arginase DNA plasmid transfection (Gotoh & Mori, 1999). In such cells, L-arginase I and II seem to

Competition between L-arginase and iNOS has also been found in activated murine macrophages incubated with another L-arginase inhibitor, nor-L-NOHA (Tenu et al., 1999). Contrarily, L-arginase induction by the type 2 cytokines IL-4 or IL-13 has been shown to inhibit macrophage NO synthesis due to increased L-arginine utilisation by L-arginase (Rutschman et al., 2001). Similar results have been obtained by using different cell types (Gotoh & Mori, 1999; Hecker et al., 1995). In -cells, both L-arginase I, the major isoform expressed in rodent pancreas, and L-arginase II, the main human isoform, seem to reciprocally regulate iNOS-dependent NO production under physiological L-arginine concentrations (Wu & Morris, 1998; Stickings et al., 2002; Castillo et al., 1993), which suggests that islet L-arginase may be able to compete with iNOS *in vivo*, where L-arginine ranges at non-saturating concentrations for both enzymes. This fact may be of relevance for -cells during Th1-driven insulitis, since L-arginine concentrations are likely to be reduced at sites of inflammation due to the release of soluble L-arginase from infiltrating macrophages (Albina et al., 1990). Corroborating this proposition is the fact that IL-1-induction of NO synthesis in RINm5F insulin secreting -like cells is accompanied by a reduced flux of L-arginine through L-arginase, an effect that appears to be mediated by L-NOHA (Cunningham et al., 1997). Hence, it is likely that, following immune cell-elicited NO production via iNOS, L-NOHA inhibits islet L-arginase activity to some degree *in vivo*, which may be strongly exacerbated by the pro-inflammatory cytokine IL-1 that inhibits L-arginase expression in -cells (Cardozo et al., 2001; Rieneck et al., 2000). In fact, a remarkable reduction in L-arginase expression has been recently observed during insulitis in

In the -cell, NH4+ may contribute to L-arginine biosynthesis, through the concerted action of carbamoyl phosphate synthetase I, ornithine transcarbamoylase, argininosuccinate synthetase and argininosuccinate lyase that produce L-arginine (**Fig. 1B).** L-Glutamate is also believed to amplify glucose-induced insulin secretion in a KATP channel-independent way (Brennan et al., 2003). However, L-glutamate is, at the same time, an obligatory substrate for GSH synthesis, which, in turn, enhances the ATP/ADP ratio by optomising mitochondrial function and scavenges ROS/RNS leading to insulin secretion. L-alanine, may replenish the -cell L-glutamate pool via an L-alanine aminotransferase-catalysed reaction. This explains why L-alanine is cytoprotective to -cells against cytokine-induced apoptosis (Cunningham et al., 2005), *i.e.*, under cytokine-stimulated NO production,

play a role in determining the route(s) for NO-elicited outcomes.

the NOD mouse model of type 1 diabetes (Rothe et al., 2002).

substrates (**Fig. 1A**) and, physiologically, L-arginine is the limiting substrate for NO production. In addition to this, pancreatic -cells express another L-arginine-metabolising enzyme, *i.e.* L-arginase (L-arginine amidinohydrolase, EC 3.5.3.1), which allows for the completion of urea production through the formation of L-ornithine and urea from L-arginine (Cunningham et al., 1997). Physiologic levels of L-arginase gene expression and activity have been measured in rat -cells and the insulin-secreting cell line RINm5F (Cunningham et al., 1997; Malaisse et al., 1989; Cardozo et al., 2001; Rieneck et el., 2000). -Cells express both the cytosolic (L-arginase I) and the mitochondrial (L-arginase II) isoforms of the enzyme. Therefore, under certain circumstances, a true competition may occur in that the activity of iNOS relative to L-arginase dictates either NO or urea production in the pancreas (compare **Fig. 1A and 1B**). Consequently, L-arginase may impair NO production by limiting the availability of L-arginine for iNOS catalysis (Wu & Morris, 1998; Boucher et al., 1999; Mori & Gotoh, 2000). This notion is supported by the finding that inhibition of L-arginase results in enhanced NO synthesis in cytokine-activated cells (Chang et al., 1998; Tenu et al., 1999).

It has been demonstrated that cytokine-elicited co-induction of both NO (iNOS) and urea (argininosuccinate synthetase and argininosuccinate lyase) metabolic pathways occurs in many cell types (Nussler et al., 1994; Hattori et al., 1994; Nagasaki et al., 1996), including -cells (Flodstrom et al., 1995), *in vitro* as well as *in vivo*. L-Arginase activity may be increased in peritoneal macrophages after exposure to LPS (Currie, 1978), while wound and peritoneal macrophages convert L-arginine to L-citrulline and L-ornithine at comparable rates, indicating that both iNOS and L-arginase pathways are functional (Granger et al., 1990). In clonal -cells, IL-1 increases L-arginase activity with concomitant increase in NO production (Cunningham et al., 1997), which suggests a kind of coordinated regulation of L-arginase and iNOS in these cells.

There is also evidence for a reciprocal regulation of NOS and L-arginase during immune responses via the antagonistic effects of cytokines released from Th1 and Th2 lymphocytes. While L-arginase activity may be induced by the "anti-inflammatory" Th2 cytokines IL-4, IL-6, IL-10, and IL-13 (Modolell et al., 1995; Waddington et al., 1998; Munder et al., 1999; Wei et al., 2000), the Th1-derived "pro-inflammatory" cytokine IFN increases iNOS expression and activity, both alone and in synergy with other pro-inflammatory cytokines, such as IL-1 and TNF (Gill et al., 1996). Reciprocal effects of Th1- and Th2-derived cytokines on L-arginase and iNOS activities have also been shown by the treatment of murine macrophages with cytokines (Modolell et al., 1995; Corraliza et al., 1995), and by coculturing murine macrophages with Th1 and Th2 T-cell clones (Munder et al., 1998). In mouse bone marrow-derived macrophages, iNOS and L-arginase activities are regulated reciprocally by Th1 and Th2 cytokines, a strategy that guarantees a precise and efficient production of NO (Modolell et al., 1995).

Because of the above statements, a Th1/Th2 lymphocyte dichotomy has been proposed to play a central role in the pathogenesis of type 1 diabetes (Rabinovitch & Suarez-Pinzon, 1998), whereas evidence suggests that the progression of the disease correlates with a Th1-type immune response (Currie, 1978; Granger et al., 1990; Simmons et al., 1996). Increased generation of NO following cytokine-elicited iNOS induction during insulitis may contribute to -cell destruction (Modolell et al., 1995; Morris et al., 1998). Therefore, competition between L-arginase and iNOS may be particularly important in protecting -cells against the establishment of type 1 diabetes.

That macrophages exposed to LPS and IFN increase iNOS expression and NO production is well known. A novel clue for the understanding of NO-mediated -cell damage is that

substrates (**Fig. 1A**) and, physiologically, L-arginine is the limiting substrate for NO production. In addition to this, pancreatic -cells express another L-arginine-metabolising enzyme, *i.e.* L-arginase (L-arginine amidinohydrolase, EC 3.5.3.1), which allows for the completion of urea production through the formation of L-ornithine and urea from L-arginine (Cunningham et al., 1997). Physiologic levels of L-arginase gene expression and activity have been measured in rat -cells and the insulin-secreting cell line RINm5F (Cunningham et al., 1997; Malaisse et al., 1989; Cardozo et al., 2001; Rieneck et el., 2000). -Cells express both the cytosolic (L-arginase I) and the mitochondrial (L-arginase II) isoforms of the enzyme. Therefore, under certain circumstances, a true competition may occur in that the activity of iNOS relative to L-arginase dictates either NO or urea production in the pancreas (compare **Fig. 1A and 1B**). Consequently, L-arginase may impair NO production by limiting the availability of L-arginine for iNOS catalysis (Wu & Morris, 1998; Boucher et al., 1999; Mori & Gotoh, 2000). This notion is supported by the finding that inhibition of L-arginase results in enhanced NO synthesis in cytokine-activated cells (Chang

It has been demonstrated that cytokine-elicited co-induction of both NO (iNOS) and urea (argininosuccinate synthetase and argininosuccinate lyase) metabolic pathways occurs in many cell types (Nussler et al., 1994; Hattori et al., 1994; Nagasaki et al., 1996), including -cells (Flodstrom et al., 1995), *in vitro* as well as *in vivo*. L-Arginase activity may be increased in peritoneal macrophages after exposure to LPS (Currie, 1978), while wound and peritoneal macrophages convert L-arginine to L-citrulline and L-ornithine at comparable rates, indicating that both iNOS and L-arginase pathways are functional (Granger et al., 1990). In clonal -cells, IL-1 increases L-arginase activity with concomitant increase in NO production (Cunningham et al., 1997), which suggests a kind of coordinated regulation of

There is also evidence for a reciprocal regulation of NOS and L-arginase during immune responses via the antagonistic effects of cytokines released from Th1 and Th2 lymphocytes. While L-arginase activity may be induced by the "anti-inflammatory" Th2 cytokines IL-4, IL-6, IL-10, and IL-13 (Modolell et al., 1995; Waddington et al., 1998; Munder et al., 1999; Wei et al., 2000), the Th1-derived "pro-inflammatory" cytokine IFN increases iNOS expression and activity, both alone and in synergy with other pro-inflammatory cytokines, such as IL-1 and TNF (Gill et al., 1996). Reciprocal effects of Th1- and Th2-derived cytokines on L-arginase and iNOS activities have also been shown by the treatment of murine macrophages with cytokines (Modolell et al., 1995; Corraliza et al., 1995), and by coculturing murine macrophages with Th1 and Th2 T-cell clones (Munder et al., 1998). In mouse bone marrow-derived macrophages, iNOS and L-arginase activities are regulated reciprocally by Th1 and Th2 cytokines, a strategy that guarantees a precise and efficient

Because of the above statements, a Th1/Th2 lymphocyte dichotomy has been proposed to play a central role in the pathogenesis of type 1 diabetes (Rabinovitch & Suarez-Pinzon, 1998), whereas evidence suggests that the progression of the disease correlates with a Th1-type immune response (Currie, 1978; Granger et al., 1990; Simmons et al., 1996). Increased generation of NO following cytokine-elicited iNOS induction during insulitis may contribute to -cell destruction (Modolell et al., 1995; Morris et al., 1998). Therefore, competition between L-arginase and iNOS may be particularly important in protecting

That macrophages exposed to LPS and IFN increase iNOS expression and NO production is well known. A novel clue for the understanding of NO-mediated -cell damage is that

et al., 1998; Tenu et al., 1999).

L-arginase and iNOS in these cells.

production of NO (Modolell et al., 1995).


*N*G-hydroxy-L-arginine (L-NOHA), an intermediate in the biosynthesis of NO, is a potent competitive inhibitor of L-arginase I (Boucher et al., 1994; Daghigh et al., 1994). Indeed, substantial amounts of this metabolite are released by LPS-treated rat alveolar macrophages (Hecker et al., 1995), while inhibition of L-arginase by L-NOHA may ensure sufficient availability of L-arginine for high-output production of NO in activated cells. L-Citrulline, the co-product of iNOS catalysis, and *S-*nitrosoglutathione (SNOG), an adduct produced by the reaction of NO with GSH, are also inhibitors of L-arginase in many cell types (Daghigh et al., 1994; Knowles & Moncada, 1994), including -cells (Cunningham et al., 1997). Hence, intermediates of NO synthesis, as well as NO itself, precisely coordinate a maximum of flux through iNOS in insulin-producing pancreatic cells (**Fig. 1**). Conversely, dexamethasone and dibutyryl cAMP block both iNOS and L-arginase expression, which is paralleled by a strong decrease of NO production (Gotoh & Mori, 1999). Additionally, macrophages treated with LPS and IFN undergo NO-dependent apoptosis, which may be prevented by L-arginase DNA plasmid transfection (Gotoh & Mori, 1999). In such cells, L-arginase I and II seem to play a role in determining the route(s) for NO-elicited outcomes.

Competition between L-arginase and iNOS has also been found in activated murine macrophages incubated with another L-arginase inhibitor, nor-L-NOHA (Tenu et al., 1999). Contrarily, L-arginase induction by the type 2 cytokines IL-4 or IL-13 has been shown to inhibit macrophage NO synthesis due to increased L-arginine utilisation by L-arginase (Rutschman et al., 2001). Similar results have been obtained by using different cell types (Gotoh & Mori, 1999; Hecker et al., 1995). In -cells, both L-arginase I, the major isoform expressed in rodent pancreas, and L-arginase II, the main human isoform, seem to reciprocally regulate iNOS-dependent NO production under physiological L-arginine concentrations (Wu & Morris, 1998; Stickings et al., 2002; Castillo et al., 1993), which suggests that islet L-arginase may be able to compete with iNOS *in vivo*, where L-arginine ranges at non-saturating concentrations for both enzymes. This fact may be of relevance for -cells during Th1-driven insulitis, since L-arginine concentrations are likely to be reduced at sites of inflammation due to the release of soluble L-arginase from infiltrating macrophages (Albina et al., 1990). Corroborating this proposition is the fact that IL-1-induction of NO synthesis in RINm5F insulin secreting -like cells is accompanied by a reduced flux of L-arginine through L-arginase, an effect that appears to be mediated by L-NOHA (Cunningham et al., 1997). Hence, it is likely that, following immune cell-elicited NO production via iNOS, L-NOHA inhibits islet L-arginase activity to some degree *in vivo*, which may be strongly exacerbated by the pro-inflammatory cytokine IL-1 that inhibits L-arginase expression in -cells (Cardozo et al., 2001; Rieneck et al., 2000). In fact, a remarkable reduction in L-arginase expression has been recently observed during insulitis in the NOD mouse model of type 1 diabetes (Rothe et al., 2002).

In the -cell, NH4+ may contribute to L-arginine biosynthesis, through the concerted action of carbamoyl phosphate synthetase I, ornithine transcarbamoylase, argininosuccinate synthetase and argininosuccinate lyase that produce L-arginine (**Fig. 1B).** L-Glutamate is also believed to amplify glucose-induced insulin secretion in a KATP channel-independent way (Brennan et al., 2003). However, L-glutamate is, at the same time, an obligatory substrate for GSH synthesis, which, in turn, enhances the ATP/ADP ratio by optomising mitochondrial function and scavenges ROS/RNS leading to insulin secretion. L-alanine, may replenish the -cell L-glutamate pool via an L-alanine aminotransferase-catalysed reaction. This explains why L-alanine is cytoprotective to -cells against cytokine-induced apoptosis (Cunningham et al., 2005), *i.e.*, under cytokine-stimulated NO production,

A Novel L-Arginine/L-Glutamine Coupling Hypothesis: Implications for Type 1 Diabetes 257

decrease in mortality due to the attenuation of pro-inflammatory type 1 cytokines (Wischmeyer et al., 2001), whereas L-arginine-enriched diet limits plasma and muscle L-glutamine depletion in head-injured rats (Moinard et al., 2006). Remarkably, however, **predominately Th1** (but not Th2) cell responses require the presence of optimal concentrations of L-glutamine (Chang et al., 1999). Since -cell death that accompanies the onset of type 1 diabetes is an essentially Th1-elicited cytotoxic challenge, it is not unreasonable to suppose that the specific recruitment of Th1 cells may greatly enhance L-glutamine and L-arginine utilisation leading to an L-arginine deficit, which causes a

The positive actions of L-arginine on viability, antioxidant status and insulin secretion are likely to reflect, in large part, the importance of GSH and the glutathione disulphide (GSSG) reductase systems as the main lines of antioxidant defence in -cells which are characterised by low levels of CAT and GSPx. In order to adequately provide GSH, -cells may either regenerate GSH from GSSG via a GSSG reductase-catalysed reaction or synthesise it, *de novo*, through the concerted action of -glutamylcysteine synthetase (-GCS) and GSH synthetase, which are ATP-consuming enzymes (see **Fig. 2** for metabolic schemes). Regeneration of GSH from GSSG, which utilises NADPH as a co-factor but does not require ATP, is metabolically less expensive than the *de novo* synthesis from the constituent amino acids (L-glutamate, L-cysteine and L-glycine). However, unlike the majority of cell types, pentose phosphate shunt activity is relatively low in -cells (Dröge, 2002), which is exacerbated by the high flux of glucose directed towards ATP production (Spinas, 1999). Therefore, -cell NADPH must be obtained from the cytosolic malic enzyme (**Fig. 2B**), capable of converting malate to pyruvate with the concomitant production of NADPH from NADP+ (MacDonald, 1995). *De novo* GSH synthesis, on the other hand, is completely dependent on the supply of L-glutamate, not only because this amino acid is a constituent of the GSH molecule, but also because L-glutamate acts as an amino acid donor in the synthesis of serine, which can subsequently, be converted to L-glycine, via a reaction requiring

We have found that L-arginine significantly increased glucose consumption in -cells, while decreasing lactate formation, regardless the presence or not of pro-inflammatory cytokines, (unpublished results, also see **Fig**. **2B**). This may suggest that L-arginine is able to divert glucose from mitochondrial CO2 production towards the formation of NADPH via the cytosolic malic enzyme so requiring that glucose-derived malate is transported from the mitochondrial matrix to the cytosol. Indeed, we believe that, in the presence of L-arginine, L-glutamate can be generated from both L-arginine and glucose (via 2-oxoglutarate formation and transamination) and is subsequently utilised for GSH synthesis (please, compare **Fig. 2B and 2C**). L-Arginine addition enhances the conversion of AMPK into its active phosphorylated form, thus favoring fatty acid oxidation and ATP synthesis while glucose metabolism is supporting malate formation and L-glutamate formation for NADPH and GSH generation respectively. This requirement, however, results in a reduction in

We have also observed that NOS-2 expression is stimulated by the cytokine cocktail (which enhances iNOS activity) but NO synthesis was not enhanced by changing L-arginine in the culture medium. This suggests that iNOS is saturated with L-arginine which, in turn, results in elevated urea production. This shunt in L-arginine metabolism efficiently preserves -cell redox status by favoring the production of GSH in conditions which generate excessive

stimulus-secretion coupling and the associated insulin release.

reduction of insulin release and redox imbalance.

tetrahydrofolate.

levels of NO (**Fig. 2C and 2D**).

L-alanine may provide L-glutamate for GSH synthesis thus avoiding oxidative stress and NO-induced apoptosis.

Since, as discussed above, -cells have poor NADPH-dependent GSSG reductase (GSRd) activity, necessary to regenerate GSH from GSSG in situations of oxidative stress, and NADPH production from the hexose monophosphate shunt is limited because -cell glycolytic activity is committed to mitochondrial ATP production during glucose-stimulated insulin release, *de novo* GSH biosynthesis from L-glutamate becomes crucial for insulin release and avoidance of -cell death. Hence, it is easy to envisage that any metabolic disequilibrium in providing L-arginine for NO-assisted insulin secretion, during secretagogue-stimulated insulin release, forces -cell metabolism to utilise L-glutamine-derived L-glutamate to synthesise GSH, thus ensuring little L-glutamate can undergo oxidative deamination via glutamate dehydrogenase (GDH) in these conditions. The kidney is considered to be the physiological producer of L-arginine since it is the only organ known to take up L-citrulline released from the metabolism of L-glutamine in the gut and release L-arginine into the blood (**Fig. 1 and 2**), although other tissues strongly express argininosuccinate synthetase and lyase but without any net delivery to the circulation (Vermeulen et al., 2007). In fasted humans, the contribution of L-glutamine via L-citrulline to the *de novo* synthesis of L-arginine is about 65% in neonates, where the gut is the major source of systemic L-arginine, even though some residual production in the adult gut could be accounted for by L-arginine release as well (Vermeulen et al., 2007). A minor part of circulating L-arginine may also be provided by the enterocyte metabolism of proline, as stated in the Introduction. Consequently, if, by any chance, the flux through the coupled L-glutamine/L-arginine pathway between intestine and kidney is reduced or lost, then the knock on consequences for NO synthesis are severe (**Fig. 1**). L-Glutamate, however, is a unique source of GSH in -cells, so that a disruption or hypofunctionality of intestinal-renal L-glutamine/L-arginine axis, would promptly decrease GSH synthesis thus reducing insulin release, leading to oxidative stress and -cell death. On the other hand, L-glutamine which is a major and immediate L-glutamate precursor, is also a primary nutrient for the maintenance of immune cell function (Curi et al., 1999; Newsholme et al., 2003; Pithon-Curi et al., 2004). Hence, we believe that an immune response triggered by an immune or chemical challenge in a redox-sensitive subject (in which the expression/activity of antioxidant and GSH enzymes is low) might decrease the availability of L-glutamine for GSH generation in -cells, leading to oxidative stress (**Fig. 1B**). Analogously, it seems likely that other situations, in which the circulating L-glutamine pool is severely endangered (Curi et al., 1999; Newsholme et al., 1987; Lagranha et al., 2008), such as in undernourishment, strenuous-exercise or cancer cachexia-associated muscle loss, chronic inflammatory diseases (including obesity), severe metabolic acidosis, major burns, polytrauma and bacteremia, should result in -cell dysfunction.

L-Glutamine deficiency can occur during periods of critical illness. In patients with catabolic diseases, plasma and muscle L-glutamine levels are dramatically reduced, which correlates with the poor prognosis and high degree of protein catabolism in those patients. For instance, in patients with major burn injury, plasma L-glutamine concentration is lower than 50% of that in normal controls and it remains low for at least 21 days after the injury (Parry-Billings et al., 1990). Conversely, in LPS-endotoxemic rats, a single dose of L-glutamine, which is known to induce anti-inflammation via HSP70 expression (Wischmeyer et al., 2003; Singleton, K.D. & Wischmeyer, P.E., 2008; Hamiel et al., 2009; Zhang et al., 2009) has been shown to attenuate the release of TNF and IL-1 and to be associated with a significant

L-alanine may provide L-glutamate for GSH synthesis thus avoiding oxidative stress and

Since, as discussed above, -cells have poor NADPH-dependent GSSG reductase (GSRd) activity, necessary to regenerate GSH from GSSG in situations of oxidative stress, and NADPH production from the hexose monophosphate shunt is limited because -cell glycolytic activity is committed to mitochondrial ATP production during glucose-stimulated insulin release, *de novo* GSH biosynthesis from L-glutamate becomes crucial for insulin release and avoidance of -cell death. Hence, it is easy to envisage that any metabolic disequilibrium in providing L-arginine for NO-assisted insulin secretion, during secretagogue-stimulated insulin release, forces -cell metabolism to utilise L-glutamine-derived L-glutamate to synthesise GSH, thus ensuring little L-glutamate can undergo oxidative deamination via glutamate dehydrogenase (GDH) in these conditions. The kidney is considered to be the physiological producer of L-arginine since it is the only organ known to take up L-citrulline released from the metabolism of L-glutamine in the gut and release L-arginine into the blood (**Fig. 1 and 2**), although other tissues strongly express argininosuccinate synthetase and lyase but without any net delivery to the circulation (Vermeulen et al., 2007). In fasted humans, the contribution of L-glutamine via L-citrulline to the *de novo* synthesis of L-arginine is about 65% in neonates, where the gut is the major source of systemic L-arginine, even though some residual production in the adult gut could be accounted for by L-arginine release as well (Vermeulen et al., 2007). A minor part of circulating L-arginine may also be provided by the enterocyte metabolism of proline, as stated in the Introduction. Consequently, if, by any chance, the flux through the coupled L-glutamine/L-arginine pathway between intestine and kidney is reduced or lost, then the knock on consequences for NO synthesis are severe (**Fig. 1**). L-Glutamate, however, is a unique source of GSH in -cells, so that a disruption or hypofunctionality of intestinal-renal L-glutamine/L-arginine axis, would promptly decrease GSH synthesis thus reducing insulin release, leading to oxidative stress and -cell death. On the other hand, L-glutamine which is a major and immediate L-glutamate precursor, is also a primary nutrient for the maintenance of immune cell function (Curi et al., 1999; Newsholme et al., 2003; Pithon-Curi et al., 2004). Hence, we believe that an immune response triggered by an immune or chemical challenge in a redox-sensitive subject (in which the expression/activity of antioxidant and GSH enzymes is low) might decrease the availability of L-glutamine for GSH generation in -cells, leading to oxidative stress (**Fig. 1B**). Analogously, it seems likely that other situations, in which the circulating L-glutamine pool is severely endangered (Curi et al., 1999; Newsholme et al., 1987; Lagranha et al., 2008), such as in undernourishment, strenuous-exercise or cancer cachexia-associated muscle loss, chronic inflammatory diseases (including obesity), severe metabolic acidosis, major burns, polytrauma and bacteremia,

L-Glutamine deficiency can occur during periods of critical illness. In patients with catabolic diseases, plasma and muscle L-glutamine levels are dramatically reduced, which correlates with the poor prognosis and high degree of protein catabolism in those patients. For instance, in patients with major burn injury, plasma L-glutamine concentration is lower than 50% of that in normal controls and it remains low for at least 21 days after the injury (Parry-Billings et al., 1990). Conversely, in LPS-endotoxemic rats, a single dose of L-glutamine, which is known to induce anti-inflammation via HSP70 expression (Wischmeyer et al., 2003; Singleton, K.D. & Wischmeyer, P.E., 2008; Hamiel et al., 2009; Zhang et al., 2009) has been shown to attenuate the release of TNF and IL-1 and to be associated with a significant

NO-induced apoptosis.

should result in -cell dysfunction.

decrease in mortality due to the attenuation of pro-inflammatory type 1 cytokines (Wischmeyer et al., 2001), whereas L-arginine-enriched diet limits plasma and muscle L-glutamine depletion in head-injured rats (Moinard et al., 2006). Remarkably, however, **predominately Th1** (but not Th2) cell responses require the presence of optimal concentrations of L-glutamine (Chang et al., 1999). Since -cell death that accompanies the onset of type 1 diabetes is an essentially Th1-elicited cytotoxic challenge, it is not unreasonable to suppose that the specific recruitment of Th1 cells may greatly enhance L-glutamine and L-arginine utilisation leading to an L-arginine deficit, which causes a reduction of insulin release and redox imbalance.

The positive actions of L-arginine on viability, antioxidant status and insulin secretion are likely to reflect, in large part, the importance of GSH and the glutathione disulphide (GSSG) reductase systems as the main lines of antioxidant defence in -cells which are characterised by low levels of CAT and GSPx. In order to adequately provide GSH, -cells may either regenerate GSH from GSSG via a GSSG reductase-catalysed reaction or synthesise it, *de novo*, through the concerted action of -glutamylcysteine synthetase (-GCS) and GSH synthetase, which are ATP-consuming enzymes (see **Fig. 2** for metabolic schemes). Regeneration of GSH from GSSG, which utilises NADPH as a co-factor but does not require ATP, is metabolically less expensive than the *de novo* synthesis from the constituent amino acids (L-glutamate, L-cysteine and L-glycine). However, unlike the majority of cell types, pentose phosphate shunt activity is relatively low in -cells (Dröge, 2002), which is exacerbated by the high flux of glucose directed towards ATP production (Spinas, 1999). Therefore, -cell NADPH must be obtained from the cytosolic malic enzyme (**Fig. 2B**), capable of converting malate to pyruvate with the concomitant production of NADPH from NADP+ (MacDonald, 1995). *De novo* GSH synthesis, on the other hand, is completely dependent on the supply of L-glutamate, not only because this amino acid is a constituent of the GSH molecule, but also because L-glutamate acts as an amino acid donor in the synthesis of serine, which can subsequently, be converted to L-glycine, via a reaction requiring tetrahydrofolate.

We have found that L-arginine significantly increased glucose consumption in -cells, while decreasing lactate formation, regardless the presence or not of pro-inflammatory cytokines, (unpublished results, also see **Fig**. **2B**). This may suggest that L-arginine is able to divert glucose from mitochondrial CO2 production towards the formation of NADPH via the cytosolic malic enzyme so requiring that glucose-derived malate is transported from the mitochondrial matrix to the cytosol. Indeed, we believe that, in the presence of L-arginine, L-glutamate can be generated from both L-arginine and glucose (via 2-oxoglutarate formation and transamination) and is subsequently utilised for GSH synthesis (please, compare **Fig. 2B and 2C**). L-Arginine addition enhances the conversion of AMPK into its active phosphorylated form, thus favoring fatty acid oxidation and ATP synthesis while glucose metabolism is supporting malate formation and L-glutamate formation for NADPH and GSH generation respectively. This requirement, however, results in a reduction in stimulus-secretion coupling and the associated insulin release.

We have also observed that NOS-2 expression is stimulated by the cytokine cocktail (which enhances iNOS activity) but NO synthesis was not enhanced by changing L-arginine in the culture medium. This suggests that iNOS is saturated with L-arginine which, in turn, results in elevated urea production. This shunt in L-arginine metabolism efficiently preserves -cell redox status by favoring the production of GSH in conditions which generate excessive levels of NO (**Fig. 2C and 2D**).

A Novel L-Arginine/L-Glutamine Coupling Hypothesis: Implications for Type 1 Diabetes 259

Fig. 2. L-Arginine-glutamate-NO coupling in -cells. Under physiological secretagoguemediated insulin release, both NO and GSH are obligatory intermediates. Accordingly, -cells have an intricate iNOS-cantered machinery to produce NO, which potentiates insulin secretion physiologically. At the same time, insulin-secreting pancreatic cells utilise glutamate-derived GSH in order to maintain redox status needed to allow hormonal secretion and to avoid a possible NO-mediated cytotoxicity. L-Arginine derived from the kidney is the physiological substrate for the NF-B-dependent iNOS-catalyzed NO production in -cells. Under

insufficient L-arginine supply, however, the high throughput of NO for -cells may be attained

Fig. 2. L-Arginine-glutamate-NO coupling in -cells. Under physiological secretagoguemediated insulin release, both NO and GSH are obligatory intermediates. Accordingly, -cells have an intricate iNOS-cantered machinery to produce NO, which potentiates insulin secretion physiologically. At the same time, insulin-secreting pancreatic cells utilise glutamate-derived GSH in order to maintain redox status needed to allow hormonal secretion and to avoid a possible NO-mediated cytotoxicity. L-Arginine derived from the kidney is the physiological substrate for the NF-B-dependent iNOS-catalyzed NO production in -cells. Under insufficient L-arginine supply, however, the high throughput of NO for -cells may be attained

A Novel L-Arginine/L-Glutamine Coupling Hypothesis: Implications for Type 1 Diabetes 263

L-Arginase is normally associated with a Km value for L-arginine that is much higher than that of iNOS but a greater Vmax value compared with iNOS (Mori, 2007), so that the Vmax/Km ratios of both enzymes are close to each other and thus these enzymes may be expected to compete for L-arginine equally in -cells. In our hands, iNOS seemed to be saturated in -cells, regardless of the presence of inflammatory cytokines, so that -cell urea production is able to furnish L-ornithine and thus L-glutamate for GSH synthesis in appropriate conditions. Moreover, L-arginine may protect -cells via the induction of haem oxygenase (HO-1) expression (data not shown). HO activity is an important detoxifying enzyme, due to its ability to scavenge haem groups thus providing redox protection (Abraham & Kappas, 2008). However, it is plausible that HO expression in -cells in response to L-arginine may also play a metabolic role, since one of its direct products, carbon monoxide (CO), has recently been reported to induce insulin secretion and to improve *in vivo* function of -cells after transplant (Abraham & Kappas, 2008). Moreover, the long-lasting expression of this enzyme has been shown to delay the progression of type 1 diabetes in NOD mice (Li et al., 2007). Hence, L-arginine can be recognised as an antioxidant in its own right, being comparable with known antioxidant stimuli, such as phytochemical supplements

Furthermore, and interestingly, chronic hyperlactataemia, in which high plasma levels of lactate block intestinal proline oxidase activity leading to severe hypocitrullinaemia and hypoargininaemia (Dillon et al., 1999), has been described as an independent risk factor for diabetes development, with lactate being an important factor for maintaining insulin resistance (DiGirolamo et al., 1992; Lovejoy et al., 1992). Conversely, L-arginine supplementation to critical care patients did induce L-glutamine rise in the plasma (Loï et al., 2009), which may be related to the fact the L-arginine supplementation spares plasma

In synthesis, L-arginine derived from the kidney (**Fig. 1**) is the physiological substrate for the NF-B-dependent iNOS-catalysed NO production in -cells. Under **insufficient** L-arginine supply, however, the high throughput of NO for -cells may be attained by the concerted action of phosphate-dependent glutaminase (GDP), glutamate dehydrogenase (GDH), aspartate aminotransferase (AsAT), carbamoylphosphate synthetase (CPS), ornithine transcarbamoylase (OTC), argininosuccinate synthetase (ASS) and argininosuccinate lyase (ASL), which, dramatically enhances the flux of glutamate towards NO production. Multiple negative feedback systems act in -cells in order to warrant L-arginine entry in iNOS metabolic pathway. This is achieved mainly due to the inhibition of L-arginase activity by L-citrulline, *N*G-hydroxy-L-arginine (L-NOHA, an intermediate in NO synthesis) and *S*-nitrosoglutathione (SNOG), which is formed during NO biosynthesis. On the other hand, -cells have to synthesise GSH from L-glutamate, L-cysteine and L-glycine, because regeneration of GSH from GSSG via NADPH-dependent GSSG reductase is relatively low in -cells because of the high flux of glucose towards ATP production that empty pentosephosphate shunt, the major NADPH-producing system. In turn, *de novo* GSH synthesis is mainly dependent on liver-derived supply of glutamate, which is not enough to allow for the enormous flux towards -glutamylcysteine synthetase and GSH synthetase in the GSH biosynthetic pathway. Therefore, muscle-derived L-alanine and L-glutamine constitute the principal sources of L-glutamate for GSH synthesis in order to spare -cell L-arginine stores. In fact, previous reports from our laboratory have highlighted the importance of L-glutamine and L-alanine for GSH generation, insulin secretion and protection against pro-

(Velmurugan et al., 2009).

glutamine pools.

by the concerted action of phosphate-dependent glutaminase (PDG), glutamate dehydrogenase (GDH), aspartate aminotransferase (not shown), carbamoylphosphate synthetase I (CPS), ornithine transcarbamoylase (OTC), argininosuccinate synthetase (ASS) and argininosuccinate lyase (ASL), which, dramatically enhances the flux of L-glutamate towards NO production. In the presence of an inflammatory NF-B-centered cytokine insult, multiple negative feedback systems act in -cells in order to warrant L-arginine entry in iNOS metabolic pathway (lower part of the figure). This is achieved mainly due to the inhibition of L-arginase activity by L-citrulline, *N*G-hydroxy- L-arginine (L-NOHA, an intermediate in NO synthesis) and *S*-nitrosoglutathione (SNOG), which is formed during NO biosynthesis. On the other hand, -cells have to synthesize GSH from L-glutamate, L-cysteine and L-glycine, once regeneration of GSH from glutathione disulphide (GSSG) via NADPH-dependent GSSG reductase is relatively low in -cells because of the high flux of glucose towards ATP production that empty pentose-phosphate shunt impairing NADPH production. In turn, *de novo* GSH synthesis is mainly dependent on liver-emanated supply of glutamate, which is not enough to allow for the enormous flux towards -glutamylcysteine synthetase (glutamatecysteine ligase) and GSH synthetase in the GSH biosynthetic pathway. Therefore, musclederived L-alanine and L-glutamine constitute the principal sources of L-glutamate for GSH synthesis. Because of this, any reduction in L-arginine supply to -cells accounts for a rapid shift in L-glutamate metabolism from GSH synthesis towards NO production. For instance, during Th1-elicited immune responses, the concerted enhancement of NF-B-mediated (\*) expression of ASS, ASL and iNOS dramatically boosts NO production from L-glutamate. If this rise in NO production is not accompanied by an enhanced L-arginine supply to -cells, NO becomes very cytotoxic. Type 2 cytokines, such as interleukin-6 (IL-6) may alleviate NO toxicity by enhancing L-arginase expression that diverts L-arginine to the formation of Lornithine and urea. At the same time, intracellular expression of the 70-kDa family of heat shock proteins (HSP70), which blocks a surplus activation of NF-B-dependent genes, is cytoprotective because it warrants an equilibrium for NO production via NF-B-dependent iNOS expression thus avoiding NO cytotoxic effects. Results from the present work reveal a novel as yet unpredicted facet of L-arginine metabolism in that an increase in its plasma concentrations (**from A to B**) could drift GSH metabolism from its original main source, via Lglutamine metabolism, towards the production of L-glutamate via the left side of the -cell urea cycle, by the consecutive action of L-arginase, pyrroline-5-carboxylate dehydrogenase (PCDH), ornithine aminotransferase (OAT), -glutamylcysteine synthetase (not shown) and GSH synthetase (not shown). Under inflammatory stimuli (**C and D**), enhancement of L-arginine concentration may alleviate the excessive flux through iNOS by limiting L-arginine availability due to its conversion into GSH. Concomitantly, elevation of L-arginine levels are thought to deviate glucose mitochondrial metabolism towards its cytosolic utilisation as a NADPH precursor via malic enzyme (ME). This favors the regeneration of more GSH molecules from GSSG under oxidative stress conditions. L-Arginine may also stimulate AMPK activation which modulates closure of KATP channels and insulin secretion. NO is also capable of activating AMPK. However, in a high L-arginine environment, the excessive activation of AMPK may stimulate lipolysis and energy saving at the expense of insulin secretion. Since physical exercise stimulates L-glutamine flux towards L-arginine production, peaks IL-6 secretion by the stretching skeletal muscle and induces HSP70 expression throughout the body tissues, exercise continues to be the cheapest and most efficient way of preventing type-1 diabetes onset. Arrow widths indicate the intensity of the metabolic flux through each pathway.

L-Arginine may also stimulate AMPK activation which modulates closure of KATP channels and insulin secretion. NO is also capable of activating AMPK. However, in a high L-arginine environment, the excessive activation of AMPK may stimulate lipolysis and energy saving at the expense of insulin secretion. Since physical exercise stimulates L-glutamine flux towards L-arginine production, peaks IL-6 secretion by the stretching skeletal muscle and induces HSP70 expression throughout the body tissues, exercise continues to be the

cheapest and most efficient way of preventing type-1 diabetes onset. Arrow widths indicate

the intensity of the metabolic flux through each pathway.

by the concerted action of phosphate-dependent glutaminase (PDG), glutamate dehydrogenase (GDH), aspartate aminotransferase (not shown), carbamoylphosphate synthetase I (CPS), ornithine transcarbamoylase (OTC), argininosuccinate synthetase (ASS) and argininosuccinate lyase (ASL), which, dramatically enhances the flux of L-glutamate towards NO production. In the presence of an inflammatory NF-B-centered cytokine insult, multiple negative feedback systems act in -cells in order to warrant L-arginine entry in iNOS metabolic pathway (lower part of the figure). This is achieved mainly due to the inhibition of L-arginase activity by L-citrulline, *N*G-hydroxy- L-arginine (L-NOHA, an intermediate in NO synthesis) and *S*-nitrosoglutathione (SNOG), which is formed during NO biosynthesis. On the other hand, -cells have to synthesize GSH from L-glutamate, L-cysteine and L-glycine, once regeneration of GSH from glutathione disulphide (GSSG) via NADPH-dependent GSSG reductase is relatively low in -cells because of the high flux of glucose towards ATP production that empty pentose-phosphate shunt impairing NADPH production. In turn, *de novo* GSH synthesis is mainly dependent on liver-emanated supply of glutamate, which is not enough to allow for the enormous flux towards -glutamylcysteine synthetase (glutamatecysteine ligase) and GSH synthetase in the GSH biosynthetic pathway. Therefore, musclederived L-alanine and L-glutamine constitute the principal sources of L-glutamate for GSH synthesis. Because of this, any reduction in L-arginine supply to -cells accounts for a rapid shift in L-glutamate metabolism from GSH synthesis towards NO production. For instance, during Th1-elicited immune responses, the concerted enhancement of NF-B-mediated (\*) expression of ASS, ASL and iNOS dramatically boosts NO production from L-glutamate. If this rise in NO production is not accompanied by an enhanced L-arginine supply to -cells, NO becomes very cytotoxic. Type 2 cytokines, such as interleukin-6 (IL-6) may alleviate NO toxicity by enhancing L-arginase expression that diverts L-arginine to the formation of Lornithine and urea. At the same time, intracellular expression of the 70-kDa family of heat shock proteins (HSP70), which blocks a surplus activation of NF-B-dependent genes, is cytoprotective because it warrants an equilibrium for NO production via NF-B-dependent iNOS expression thus avoiding NO cytotoxic effects. Results from the present work reveal a novel as yet unpredicted facet of L-arginine metabolism in that an increase in its plasma concentrations (**from A to B**) could drift GSH metabolism from its original main source, via Lglutamine metabolism, towards the production of L-glutamate via the left side of the -cell urea cycle, by the consecutive action of L-arginase, pyrroline-5-carboxylate dehydrogenase (PCDH), ornithine aminotransferase (OAT), -glutamylcysteine synthetase (not shown) and GSH synthetase (not shown). Under inflammatory stimuli (**C and D**), enhancement of L-arginine concentration may alleviate the excessive flux through iNOS by limiting L-arginine availability due to its conversion into GSH. Concomitantly, elevation of L-arginine levels are thought to deviate glucose mitochondrial metabolism towards its cytosolic utilisation as a NADPH precursor via malic enzyme (ME). This favors the regeneration of more GSH molecules from GSSG under oxidative stress conditions.

L-Arginase is normally associated with a Km value for L-arginine that is much higher than that of iNOS but a greater Vmax value compared with iNOS (Mori, 2007), so that the Vmax/Km ratios of both enzymes are close to each other and thus these enzymes may be expected to compete for L-arginine equally in -cells. In our hands, iNOS seemed to be saturated in -cells, regardless of the presence of inflammatory cytokines, so that -cell urea production is able to furnish L-ornithine and thus L-glutamate for GSH synthesis in appropriate conditions. Moreover, L-arginine may protect -cells via the induction of haem oxygenase (HO-1) expression (data not shown). HO activity is an important detoxifying enzyme, due to its ability to scavenge haem groups thus providing redox protection (Abraham & Kappas, 2008). However, it is plausible that HO expression in -cells in response to L-arginine may also play a metabolic role, since one of its direct products, carbon monoxide (CO), has recently been reported to induce insulin secretion and to improve *in vivo* function of -cells after transplant (Abraham & Kappas, 2008). Moreover, the long-lasting expression of this enzyme has been shown to delay the progression of type 1 diabetes in NOD mice (Li et al., 2007). Hence, L-arginine can be recognised as an antioxidant in its own right, being comparable with known antioxidant stimuli, such as phytochemical supplements (Velmurugan et al., 2009).

Furthermore, and interestingly, chronic hyperlactataemia, in which high plasma levels of lactate block intestinal proline oxidase activity leading to severe hypocitrullinaemia and hypoargininaemia (Dillon et al., 1999), has been described as an independent risk factor for diabetes development, with lactate being an important factor for maintaining insulin resistance (DiGirolamo et al., 1992; Lovejoy et al., 1992). Conversely, L-arginine supplementation to critical care patients did induce L-glutamine rise in the plasma (Loï et al., 2009), which may be related to the fact the L-arginine supplementation spares plasma glutamine pools.

In synthesis, L-arginine derived from the kidney (**Fig. 1**) is the physiological substrate for the NF-B-dependent iNOS-catalysed NO production in -cells. Under **insufficient** L-arginine supply, however, the high throughput of NO for -cells may be attained by the concerted action of phosphate-dependent glutaminase (GDP), glutamate dehydrogenase (GDH), aspartate aminotransferase (AsAT), carbamoylphosphate synthetase (CPS), ornithine transcarbamoylase (OTC), argininosuccinate synthetase (ASS) and argininosuccinate lyase (ASL), which, dramatically enhances the flux of glutamate towards NO production. Multiple negative feedback systems act in -cells in order to warrant L-arginine entry in iNOS metabolic pathway. This is achieved mainly due to the inhibition of L-arginase activity by L-citrulline, *N*G-hydroxy-L-arginine (L-NOHA, an intermediate in NO synthesis) and *S*-nitrosoglutathione (SNOG), which is formed during NO biosynthesis. On the other hand, -cells have to synthesise GSH from L-glutamate, L-cysteine and L-glycine, because regeneration of GSH from GSSG via NADPH-dependent GSSG reductase is relatively low in -cells because of the high flux of glucose towards ATP production that empty pentosephosphate shunt, the major NADPH-producing system. In turn, *de novo* GSH synthesis is mainly dependent on liver-derived supply of glutamate, which is not enough to allow for the enormous flux towards -glutamylcysteine synthetase and GSH synthetase in the GSH biosynthetic pathway. Therefore, muscle-derived L-alanine and L-glutamine constitute the principal sources of L-glutamate for GSH synthesis in order to spare -cell L-arginine stores. In fact, previous reports from our laboratory have highlighted the importance of L-glutamine and L-alanine for GSH generation, insulin secretion and protection against pro-

A Novel L-Arginine/L-Glutamine Coupling Hypothesis: Implications for Type 1 Diabetes 265

Fig. 3. Psychological stress and autoimmune diabetes. Different stressful situations may lead to the activation of sympathetic-corticotropin-releasing hormone (CRH)-histamine axis that triggers a Th1-specific immunoinflammatory response. Peripheral sympathetic nervederived CRH released under acute psychological stressful situations is capable of

stimulating mast cells and Th1 lymphocytes, which arm an immunoinflammatory response. Auto-reactive Th1 cell subset and its cytokine products (type 1 cytokines, T1-CK) raised against islet -cell antigen(s) mediate the activation of macrophages and Th1 lymphocytes, favouring insulitis. Additionally, other predisposing factors may also exacerbate -cell

levels (Wilmore, 2004). Curiously, intraperitoneal L-arginine injection, where the physiological coupling of L-glutamine/L-arginine through the intestinal-renal axis is bypassed, does **not** improve diabetes in animal models. On the contrary, it seems to worsen it (Mohan & Das, 1998), while **oral** administration of L-arginine to alloxan-treated rats restores blood glucose and insulin levels (Vasilijevic et al., 2007). Oral L-arginine administration has also been shown to improve, but not completely, peripheral and hepatic insulin sensitivity in type 2 diabetes (Piatti et al., 2001), where oxidative stress (Carvalho-Filho et al., 2005; Oliveira et al., 2003; Hirabara et al., 2006) and NO overproduction (Newsholme et al., 2007; Carvalho-Filho et al., 2005) are also involved. If this is so, nutritional management of L-glutamine and/or L-arginine, **enterally** administered in order to allow for the physiological re-establishment of L-glutamine/L-arginine homeostasis (Vermeulen et al., 2007), may rescue -cell redox balance in ongoing type 1 diabetes. Additionally, skeletal muscle is a major site for L-glutamine synthesis in the human body and contains over 90% of the whole-body L-glutamine pool. Quantitative studies in humans

injury and the onset of type 1 diabetes mellitus (T1DM).

inflammatory cytokines (Brennan et al., 2003; Brennan et a., 2002; Cunninham et al., 2005). Because of this, **any reduction** in L-arginine supply to -cells accounts for a rapid shift in L-glutamate metabolism from GSH synthesis towards NO production. For instance, during Th1-elicited immune responses (*e.g.* as **in Fig. 2C and 2D**), the concerted enhancement of nuclear factor NF-B-mediated expression of ASS, ASL and iNOS dramatically boosts NO production from L-glutamate. If this rise in NO production is not accompanied by an enhanced L-arginine supply to -cells, NO becomes very cytotoxic. Type 2 cytokines (T2-CK) may alleviate NO toxicity by enhancing L-arginase expression that deviates L-arginine to the formation of L-ornithine and urea.

## **6. Psychological stress and the role peripheral sympathetic nervous systemhistamine-CRH axis activation in type 1 diabetes**

It has long been recognised that stressful situations are closely related to the onset of type 1 diabetes. In fact, many stressful conditions that are associated with immune system imbalances, including psychological ones, are associated with the incidence of type 1 diabetes (Soltesz, 2003; Dahlquist, 2006). Indeed, it has recently been shown that stressful life events and psychological dysfunctions dramatically augment the likelihood of the incidence of type 1 diabetes in children and adolescents (Sipetic et al., 2007). These include parents' job-related changes or lost job, severe accidents, hospitalization or death of a close friend, quarrels between parents, war, near-drowning in a pool, falling down, being an unhurt participant of an accident, conflicts with parents/teacher/neighbours, to be lost in town, physical attack, failure in competition, penalty, examination, death of pet, presence of lightning strike, loss of housing accommodation and learning problems.

As a general rule, stress is considered as immunosuppressive. Surprisingly, however, a growing body of evidence strongly suggests that acute stress serves as a pro-inflammatory stimulus via the production of corticotropin-releasing hormone (CRH) by peripheral sympathetic nerve terminals (Elenkov et al., 1999). CRH stimulates lymphocyte proliferation (McGillis et al., 1989; Jessop et al., 1997) and secretion of IL-1 and IL-2 by mononuclear cells isolated from the peripheral blood of healthy subjects (Singh & Leu, 1990). Peripheral CRH exerts a pro-inflammatory effect in autoimmune diseases with a selective increase in Th1-type responses, which is mediated by an NF-B-dependent pathway (Benou et al., 2005). Additionally, it is possible that, upon a stressful situation, peripherally delivered CRH activates mast cells that secrete histamine, which acts via H1 receptors to induce local inflammation (Elenkov et al., 1999). In fact, diabetes is associated with increased basal hypothalamus-pituitary-adrenal (HPA) activity and impaired stress responsiveness (Chan et al., 2005). Therefore, psychological stress may selectively activate Th1 lymphocytes that mediate type-1 cytokine-induced iNOS expression, exacerbated NO production and -cell cytotoxicity. Enhanced Th1 activity, in turn, increases L-glutamine utilisation with the consequent shift of L-glutamate metabolism from GSH biosynthesis towards NO production, as discussed above (Fig. 2 and 3).

Taken together, these findings suggest that psychological stress may have a dual and crosspotentiating role in determining the onset of type 1 diabetes: an immunoinflammatory (Fig. 3) and a metabolic one (Fig. 2C and 2D). Arguing in proof of such a hypothesis is the observation that orally administered L-arginine supplementation significantly improves patient status in a series of different pathological conditions associated with immune dysfunctions, including in pre-term neonates (Wu et al., 2004), without increasing urea

inflammatory cytokines (Brennan et al., 2003; Brennan et a., 2002; Cunninham et al., 2005). Because of this, **any reduction** in L-arginine supply to -cells accounts for a rapid shift in L-glutamate metabolism from GSH synthesis towards NO production. For instance, during Th1-elicited immune responses (*e.g.* as **in Fig. 2C and 2D**), the concerted enhancement of nuclear factor NF-B-mediated expression of ASS, ASL and iNOS dramatically boosts NO production from L-glutamate. If this rise in NO production is not accompanied by an enhanced L-arginine supply to -cells, NO becomes very cytotoxic. Type 2 cytokines (T2-CK) may alleviate NO toxicity by enhancing L-arginase expression that deviates L-arginine to the

**6. Psychological stress and the role peripheral sympathetic nervous system-**

It has long been recognised that stressful situations are closely related to the onset of type 1 diabetes. In fact, many stressful conditions that are associated with immune system imbalances, including psychological ones, are associated with the incidence of type 1 diabetes (Soltesz, 2003; Dahlquist, 2006). Indeed, it has recently been shown that stressful life events and psychological dysfunctions dramatically augment the likelihood of the incidence of type 1 diabetes in children and adolescents (Sipetic et al., 2007). These include parents' job-related changes or lost job, severe accidents, hospitalization or death of a close friend, quarrels between parents, war, near-drowning in a pool, falling down, being an unhurt participant of an accident, conflicts with parents/teacher/neighbours, to be lost in town, physical attack, failure in competition, penalty, examination, death of pet, presence of

As a general rule, stress is considered as immunosuppressive. Surprisingly, however, a growing body of evidence strongly suggests that acute stress serves as a pro-inflammatory stimulus via the production of corticotropin-releasing hormone (CRH) by peripheral sympathetic nerve terminals (Elenkov et al., 1999). CRH stimulates lymphocyte proliferation (McGillis et al., 1989; Jessop et al., 1997) and secretion of IL-1 and IL-2 by mononuclear cells isolated from the peripheral blood of healthy subjects (Singh & Leu, 1990). Peripheral CRH exerts a pro-inflammatory effect in autoimmune diseases with a selective increase in Th1-type responses, which is mediated by an NF-B-dependent pathway (Benou et al., 2005). Additionally, it is possible that, upon a stressful situation, peripherally delivered CRH activates mast cells that secrete histamine, which acts via H1 receptors to induce local inflammation (Elenkov et al., 1999). In fact, diabetes is associated with increased basal hypothalamus-pituitary-adrenal (HPA) activity and impaired stress responsiveness (Chan et al., 2005). Therefore, psychological stress may selectively activate Th1 lymphocytes that mediate type-1 cytokine-induced iNOS expression, exacerbated NO production and -cell cytotoxicity. Enhanced Th1 activity, in turn, increases L-glutamine utilisation with the consequent shift of L-glutamate metabolism from GSH biosynthesis towards NO

Taken together, these findings suggest that psychological stress may have a dual and crosspotentiating role in determining the onset of type 1 diabetes: an immunoinflammatory (Fig. 3) and a metabolic one (Fig. 2C and 2D). Arguing in proof of such a hypothesis is the observation that orally administered L-arginine supplementation significantly improves patient status in a series of different pathological conditions associated with immune dysfunctions, including in pre-term neonates (Wu et al., 2004), without increasing urea

formation of L-ornithine and urea.

**histamine-CRH axis activation in type 1 diabetes** 

production, as discussed above (Fig. 2 and 3).

lightning strike, loss of housing accommodation and learning problems.

Fig. 3. Psychological stress and autoimmune diabetes. Different stressful situations may lead to the activation of sympathetic-corticotropin-releasing hormone (CRH)-histamine axis that triggers a Th1-specific immunoinflammatory response. Peripheral sympathetic nervederived CRH released under acute psychological stressful situations is capable of stimulating mast cells and Th1 lymphocytes, which arm an immunoinflammatory response. Auto-reactive Th1 cell subset and its cytokine products (type 1 cytokines, T1-CK) raised against islet -cell antigen(s) mediate the activation of macrophages and Th1 lymphocytes, favouring insulitis. Additionally, other predisposing factors may also exacerbate -cell injury and the onset of type 1 diabetes mellitus (T1DM).

levels (Wilmore, 2004). Curiously, intraperitoneal L-arginine injection, where the physiological coupling of L-glutamine/L-arginine through the intestinal-renal axis is bypassed, does **not** improve diabetes in animal models. On the contrary, it seems to worsen it (Mohan & Das, 1998), while **oral** administration of L-arginine to alloxan-treated rats restores blood glucose and insulin levels (Vasilijevic et al., 2007). Oral L-arginine administration has also been shown to improve, but not completely, peripheral and hepatic insulin sensitivity in type 2 diabetes (Piatti et al., 2001), where oxidative stress (Carvalho-Filho et al., 2005; Oliveira et al., 2003; Hirabara et al., 2006) and NO overproduction (Newsholme et al., 2007; Carvalho-Filho et al., 2005) are also involved. If this is so, nutritional management of L-glutamine and/or L-arginine, **enterally** administered in order to allow for the physiological re-establishment of L-glutamine/L-arginine homeostasis (Vermeulen et al., 2007), may rescue -cell redox balance in ongoing type 1 diabetes. Additionally, skeletal muscle is a major site for L-glutamine synthesis in the human body and contains over 90% of the whole-body L-glutamine pool. Quantitative studies in humans

A Novel L-Arginine/L-Glutamine Coupling Hypothesis: Implications for Type 1 Diabetes 267

are currently evaluating the effects of acute and chronic (training) exercise sessions (swimming) on HSP70 pathways and L-glutamine/L-arginine coupling enzymes in

Continued supply of L-arginine, physiologically provided by the metabolism of L-glutamine via the intestinal-renal axis and from skeletal muscle, which is enhanced during exercise, is essential for -cell functional integrity and, indeed, for -cell defence. The dysregulation of immune system function, characteristic of Th1-elicited -cell toxicity and impaired insulin secretion, which accompany the onset of type 1 diabetes, may be triggered when an individual faces a strong **psychological stress** that determines an enhanced L-glutamine utilisation by Th1 lymphocytes. The oxidative stress that takes place upon reduced intracellular GSH levels allows for the activation of NF-B, which, in turn, positively feeds back on iNOS expression and activity, thus perpetuating the inflammatory process within -cells where **excess** NO is harmful. Defective HSP70 induction in response to physiological levels of intraislet NO may also be involved in the pathogenesis of type 1 diabetes. Physical exercise, on the other hand, is capable of inducing a huge production and release of IL-6, which is a key anti-inflammatory mediator that suppresses NF-B-dependent responses. Moreover, exercise-elicited activation of HSP70 biochemical pathways completely blocks NF-B activation, impedes apoptosis and is cytoprotective due to HSP70 chaperone activity, which protects against protein denaturation. HSP70 induction is also associated with enhanced Th2 cell activity over Th1. Metabolically, exercise may restore L-glutamine supply thus normalizing pancreatic production of NO from kidney-derived L-arginine, and not from L-glutamate which is necessary for GSH synthesis and antioxidant defence. Thus the enormous changes in human life style, compared with that of our 3-4 million-old ancestors, could be related with our current inability in maintaining healthy -cells. As previously argued (Krause & Homem de Bittencourt, 2008), we advocate that present-day levels of physical activity and dietary patterns (Simopoulos, 2006; Wisloff et al., 2005) seem to have changed much faster than the time needed to allow evolutionary metabolic changes. In other words, our metabolism evolved to fit a level of physical activity and availability of a variety of food supplies different from those of nowadays (favouring energy conservation and storage). As a corollary, unless humans enhance their pattern of physical activity, diabetes will become more and more of a risk factor in the population. Therefore, the notion that -cells are solely bystanders of oxidative stress-mediated cell toxicity because their antioxidant defences fail in managing physiological stress is an unfortunate misconception. Since the L-glutamine/L-arginine duet may influence -cell function and survival, the knowledge of physiologically adequate levels and fluxes of both amino acids may serve as a predictor of -cell susceptibility to dysfunction or death in diabetes. Additionally, although the possibility of pharmacologically exploiting Th1/Th2 duality relative to L-arginine metabolism may open new avenues for diabetes therapeutics, physical exercise is still the cheapest and easiest physiological measure to avoid the onset and/or worsening of diabetes. In summary, if the prevention of diabetes is dependent on HSP70 expression and both restoration of adequate L-arginine supply to -cells and blockage of NF-B overstimulation, moderate physical exercise is presented as the most convenient solution for

animal pancreatic islets and isolated -cells.

**8. Conclusion** 

these two lacunes.

(Newsholme et al., 2003) have demonstrated that, in the postabsorptive state, 60% of the amino acids released comprise L-alanine plus L-glutamine (**Fig. 1A**). Therefore, moderate physical exercise, which is known to accelerate the rate of L-glutamine delivery into the circulation, may be of value in protecting L-glutamine/L-arginine metabolic coupling between the gut and -cells.
