Alloxan and Streptozotocin Diabetes - Sigurd Lenzen
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Sigurd Lenzen
Alloxan and Streptozotocin Diabetes
Abstract While the mechanisms of cytotoxic action of the two
compounds are different, the mechanism of the selectivity
The cytotoxic glucose analogues alloxan and streptozoto- of the beta cell action is identical.
cin are the most prominent diabetogenic chemical agents In 1838 WoÈhler and Liebig synthesized a pyrimidine
in experimental diabetes research. While the mechanism derivative (WoÈhler and Liebig, 1838), which they later
of the selectivity of pancreatic beta cell toxicity is identi- called alloxan (for review see: Lenzen and Panten,
cal the mechanisms of the cytotoxic action of the two 1988a; Lenzen et al., 1996b). In 1943, alloxan became
compounds are different. Both are selectively toxic to of interest in diabetes research when Shaw Dunn and
beta cells because they preferentially accumulate in beta McLetchie reported that this compound could induce dia-
cells as glucose analogues through uptake via the betes (Dunn, 1943; Dunn and McLetchie, 1943) due to a
GLUT2 glucose transporter. Alloxan, in the presence of specific necrosis of the pancreatic beta cells in experi-
intracellular thiols, especially glutathione, generates re- mental animals (Dunn et al., 1943a,b; JoÈrns et al., 1997;
active oxygen species (ROS) in a cyclic reaction with Peschke et al., 2000). The emerging insulinopenia causes
its reduction product, dialuric acid. The beta cell toxic a state of experimental diabetes mellitus called ªalloxan
action of alloxan is initiated by free radicals formed in diabetesº (Dunn and McLetchie, 1943; Goldner and Go-
this redox reaction. Autoxidation of dialuric acid gener- mori, 1943; McLetchie, 1982). The reduction product of
ates superoxide radicals (O·2 ±), hydrogen peroxide alloxan, dialuric acid (WoÈhler and Liebig, 1838), has also
(H2O2) and, in a final iron catalysed reaction step, hy- been shown to be diabetogenic in experimental animals
droxyl radicals (OH·). These hydroxyl radicals are ulti- (BruÈckmann and Wertheimer, 1945, 1947; Bailey et al.,
mately responsible for the death of the beta cells with 1946; Saviano and De Franciscis, 1946; Merlini, 1951),
their particularly low antioxidative defence capacity and to cause, like lipophilic alloxan derivatives such as
and the ensuing state of insulin-dependent ªalloxan dia- butylalloxan, identical ultrastructural changes (JoÈrns
betesº. As a thiol reagent, alloxan selectively inhibits et al., 1997).
glucose-induced insulin secretion through its ability to Streptozotocin is an antimicrobial agent originally de-
specifically inhibit the glucokinase through oxidation of rived from the soil micro-organism Streptomyces achro-
functionally essential thiol groups in this protein, there- mogenes (White, 1963; Herr et al., 1967; Weiss, 1982).
by impairing oxidative metabolism and the glucose sen- It has been used as a chemotherapeutic alkylating agent
sor function of this signalling enzyme of the beta cell. in the treatment of metastasizing pancreatic islet cell tu-
Streptozotocin, after uptake into the beta cells, is split mours and in other malignancies (Evans et al., 1965;
up into its glucose and methylnitrosourea moiety. The Schein et al., 1974). In 1963 Rakieten and collaborators
latter, due to its alkylating properties, modifies biological reported that streptozotocin is diabetogenic (Rakieten
macromolecules, fragments DNA and destroys the beta et al., 1963). Ever since streptozotocin has been the agent
cells, causing a state of insulin-dependent ªstreptozoto- of choice to induce experimental diabetes mellitus in lab-
cin diabetesº. The ability of streptozotocin to inhibit oratory animals (Rerup, 1970). This insulinopenia syn-
glucose-induced insulin secretion, likewise, can be ex- drome, called ªstreptozotocin diabetesº (Schein et al.,
plained with its alkylating potency, targeting mitochon- 1967), is due to the ability of this compound to induce a
drial DNA and thereby impairing the signalling function specific necrosis of the pancreatic beta cells (Arison
of beta cell mitochondrial metabolism. et al., 1967; Lenzen et al., 1996b).
After decades of research a unifying explanation for
the molecular mechanism of toxic action of these two
most prominent diabetogenic agents can be provided. As
Introduction the understanding of the chemical reactivity of these
compounds is crucial for the understanding of their diabe-
Alloxan and streptozotocin are the most prominent diabe- togenicity, this review will provide an integrated view
togenic chemical compounds in experimental diabetes re- upon the chemical properties and the biological effects
search. Both compounds are cytotoxic glucose analogues. of alloxan and streptozotocin. It is not the aim to provide120 Sigurd Lenzen
an extensive review on the entire literature on these two
diabetogenic agents, since this has been covered in pre-
vious reviews (Rerup, 1970; Cooperstein and Watkins,
1981; Weiss, 1982; Preston, 1985; Lenzen and Panten,
1988a; Szkudelski, 2001; BroÈmme and Peschke, 2003).
Alloxan diabetes and streptozotocin diabetes
Alloxan and streptozotocin, when injected into to an ex-
perimental animal, induce a multiphasic blood glucose re-
sponse (Fig. 1), which is accompanied by corresponding
inverse changes in the plasma insulin concentration
(Goldner and Gomori, 1944; Shipley and Beyer, 1947;
Wrenshall et al., 1950; Suchowsky, 1953; Lundquist and
Rerup, 1967; Boquist, 1968a,b; Rerup and Tarding,
1969; Rerup, 1970; Losert et al., 1971; Schusdziarra
et al., 1979; Tasaka et al., 1988; West et al., 1996; Szku-
delski, 2001) as well as sequential ultrastructural beta cell Fig. 1: Phasic blood glucose response to a diabetogenic dose of
changes finally leading to necrotic cell death (Wellmann alloxan (tetraphasic; I±IV) or streptozotocin (triphasic; first
et al., 1967; Boquist, 1968a,b, 1977; Richter et al., 1971; phase does not develop in the case of streptozotocin; II±IV).
Tomita, 1972; Boquist and Lorentzon, 1979; Lorentzon
and Boquist, 1979; Lenzen et al., 1996b; JoÈrns et al., This 2nd phase of the blood glucose response starts
1997; Mythili et al., 2004). with a rise of the blood glucose concentration one hour
A 1st transient hypoglycaemic phase lasting maximally after administration of the toxins, while at the same time
30 min starts within the first minutes after alloxan injec- the plasma insulin concentration decreases. This is the
tion (Wrenshall et al., 1950; Suchowsky, 1953; Tasaka first hyperglycaemic phase after the first contact of the
et al., 1988). This short-lived hypoglycaemic response is beta cells with the toxins (Goldner and Gomori, 1944;
the result of a transient stimulation of insulin secretion Shipley and Beyer, 1947; Wrenshall et al., 1950; Su-
(Pagliara et al., 1977; Weaver et al., 1978c; Goto et al., chowsky, 1953; Lundquist and Rerup, 1967; Boquist,
1980; Kliber et al., 1996), as documented by an increase 1968a,b; Rerup and Tarding, 1969; Rerup, 1970; Losert
of the plasma insulin concentration (Tasaka et al., 1988). et al., 1971; Schusdziarra et al., 1979; Tasaka et al.,
The underlying mechanism of this transient hyperinsu- 1988; West et al., 1996). This pronounced hyperglycae-
linaemia is a temporary increase of ATP availability due mia usually lasts 2±4 hours and is accompanied by de-
to inhibition of glucose phosphorylation through glucoki- creased plasma insulin concentrations (Lundquist and
nase inhibition. This initial transient hypoglycaemic Rerup, 1967; Tasaka et al., 1988). These changes result
phase is not observed in the case of a streptozotocin in- from an inhibition of insulin secretion from the pancreatic
jection because, at variance from alloxan, streptozotocin beta cells, which both toxins induce due to their beta cell
does not inhibit glucokinase (Lenzen et al., 1987b). Dur- toxicity. Morphologically this phase is characterized in
ing these first five minutes after toxin exposure the beta the beta cells by intracellular vacuolization, dilatation of
cells show no morphological signs of damage (JoÈrns the cisternae of the rough endoplasmic reticulum, a de-
et al., 1997). Up to one hour after exposure to both alloxan creased Golgi area, a reduction of insulin content and of
and streptozotocin, at a time, when normoglycaemia and the number of secretory granules as well as swollen mito-
normoinsulinaemia are prevailing, morphological beta chondria with a loss of cristae of the inner membranes
cell changes are still minimal. Discrete changes in the (Wellmann et al., 1967; Boquist, 1968a,b, 1977; Richter
form of a pale cytoplasm and beginning intracellular mi- et al., 1971; Tomita, 1972; Boquist and Lorentzon, 1979;
crovacuolisation, dilatation of the cisternae of the rough Lorentzon and Boquist, 1979; Lenzen et al., 1996b; JoÈrns
endoplasmic reticulum and mildly swollen mitochondria et al., 1997). The observations are consistent with a de-
become visible after 20±30 min, apparent in particular fective mitochondrial energy production (Boquist, 1977;
on the ultrastructural level (Wellmann et al., 1967; Bo- Boquist and Lorentzon, 1979; Lorentzon and Boquist,
quist, 1968a,b, 1977; Richter et al., 1971; Tomita, 1972; 1979) due to the toxic action of the diabetogens and an
Boquist and Lorentzon, 1979; Lorentzon and Boquist, inhibition of pre-proinsulin biosynthesis as well as pro-
1979; Lenzen et al., 1996b; JoÈrns et al., 1997). At this cessing, packaging and storage of insulin in the secretory
stage the secretory granules are still normal in number granules (Jain and Logothetopoulos, 1976; Niki et al.,
and ultrastructure. These early signs can be considered 1976). This is the underlying cause for the inhibition of
potentially reversible intracellular lesions. insulin secretion (Lenzen and Panten, 1988a,b) and theAlloxan and Streptozotocin Diabetes 121 resultant hyperglycaemia and hypoinsulinaemia (Lund- (Lenzen et al., 1996b; JoÈrns et al., 1997), proving the beta quist and Rerup, 1967; Tasaka et al., 1988) during this cell selective character of the toxic action of these diabe- phase. togenic compounds. Cell debris of the dying beta cells is The 3rd phase of the blood glucose response is again a removed by scavenger macrophages, which are not acti- hypoglycaemic phase typically 4±8 hours after the injec- vated (Williamson and Lacy, 1959). After destruction of tion of the toxins, which lasts several hours (Jacobs, the beta cells, when survival through insulin supplemen- 1937; Goldner and Gomori, 1944; Shipley and Beyer, tation is secured, so-called end stage islets reside in the 1947; Wrenshall et al., 1950; Suchowsky, 1953; Lund- pancreas, exclusively composed of non-beta cells (JoÈrns quist and Rerup, 1967; Boquist, 1968a,b; Rerup and Tard- et al., 2001). ing, 1969; Rerup, 1970; Losert et al., 1971; Schusdziarra Thus, injections of alloxan and streptozotocin induce et al., 1979; Tasaka et al., 1988; West et al., 1996). It may principally the same blood glucose and plasma insulin re- be so severe that it causes convulsions (Goldner and Go- sponse and induce an insulin-dependent type 1 like dia- mori, 1944) and may even be fatal without glucose ad- betes syndrome. All morphological features of beta cell ministration (Shipley and Beyer, 1947; Losert et al., destruction are characteristic for a necrotic cell death 1971), in particular when liver glycogen stores are de- (Lenzen et al., 1996b; JoÈrns et al., 1997; Peschke et al., pleted through starvation (Wrenshall et al., 1950). This 2000). This is clearly at variance from the situation of severe transitional hypoglycaemia is the result of a flood- autoimmune type 1 diabetes mellitus, both in humans as ing of the circulation with insulin as a result of the toxin- well as in mouse and rat models of the disease, where induced secretory granule and cell membrane rupture beta cell demise is the result of apoptotic cell death with- (Banerjee, 1945, 1948). Pancreatectomy prevents this out leakage of insulin from ruptured secretory granules phase (Griffiths, 1948). As documented through morpho- (Lenzen et al., 2001; Lally and Bone, 2002; JoÈrns et al., logical and ultrastructural analyses (Wellmann et al., 2004, 2005). 1967; Boquist, 1968a,b, 1977; Richter et al., 1971; Tomi- ta, 1972; Boquist and Lorentzon, 1979; Lorentzon and Boquist, 1979; Lenzen et al., 1996b; JoÈrns et al., 1997; Alloxan mechanism of action Mythili et al., 2004), these changes comprise not only ruptures of the secretory granules with loss of their insu- The diabetogenic agent alloxan (Fig. 2) has two distinct lin content and of the plasma membrane but also of other pathological effects interfering with the physiological subcellular organelles, including the cisternae of the function of the pancreatic beta cells. It selectively inhibits rough endoplasmic reticulum and the Golgi complex. glucose-induced insulin secretion through its ability to The outer and inner membranes of the mitochondria also specifically inhibit the glucokinase, the glucose sensor of loose their structural integrity in this phase. Expression of the beta cell, and it causes a state of insulin-dependent functionally essential proteins such as GLUT2 glucose diabetes mellitus through its ability to induce a selective transporter and glucokinase is lost and insulin protein is necrosis of the beta cells. These two effects of alloxan undetectable. The beta cell nuclei are shrunken with con- can be assigned to specific chemical properties of allox- densed chromatin as signs of pyknosis. Nuclei show no an. The common denominator of these effects is selective TUNEL positive staining. These changes are irreversible cellular uptake and accumulation of alloxan by the beta and highly characteristic for a necrotic type of beta cell cell. death. The 4th phase of the blood glucose response is the final permanent diabetic hyperglycaemic phase, which charac- terizes alloxan and streptozotocin diabetes (Goldner and Gomori, 1944; Shipley and Beyer, 1947; Wrenshall et al., 1950; Suchowsky, 1953; Lundquist and Rerup, 1967; Bo- quist, 1968a,b; Rerup and Tarding, 1969; Rerup, 1970; Losert et al., 1971; Schusdziarra et al., 1979; Tasaka Fig. 2: Chemical formulas of alloxan, dialuric acid, and butyl- et al., 1988; West et al., 1996). Morphological and ultra- alloxan. structural analyses document a complete degranulation and loss of the integrity of the beta cells within 12±24± 48 h after administration of the toxins (Wellmann et al., Chemical properties of alloxan 1967; Boquist, 1968a,b, 1977; Richter et al., 1971; Tomi- ta, 1972; Boquist and Lorentzon, 1979; Lorentzon and Alloxan (2,4,5,6-tetraoxypyrimidine; 2,4,5,6-pyrimidine- Boquist, 1979; Lenzen et al., 1996b; JoÈrns et al., 1997; tetrone) is: Mythili et al., 2004). Non-beta cells such as alpha cells ± An oxygenated pyrimidine derivative. It is also a barbi- and other endocrine and non-endocrine islet cell types as turic acid derivative (5-ketobarbituric acid). well as the extrapancreatic parenchyma remain intact ± Present as alloxan hydrate in aqueous solution.
122 Sigurd Lenzen
± A beta cell toxic glucose analogue with a molecular can enter the beta cells through this protein pore unre-
shape similar to that of glucose (Weaver et al., 1979). stricted, where it is accumulated selectively (HammarstroÈm
± A very hydrophilic compound (partition coefficient et al., 1967; Boquist et al., 1983; Malaisse et al., 2001)
±1.8) (Lenzen and Munday, 1991). and executes its biological effects.
± A week acid (Patterson et al., 1949).
± Chemically instable in buffer solutions with a half-life
of 1.5 min at pH 7.4 and 37 °C decomposing to alloxanic Biological effects of alloxan due to thiol group
acid (Patterson et al., 1949; Lenzen and Munday, 1991). reactivity: Selective inhibition of glucose-induced
± Stable at acid pH (0.01 M HCl) (Lenzen and Munday, insulin secretion through inhibition of glucokinase
1991).
± For experimentation concentrated stock solutions in The selective inhibition of glucose-induced insulin secre-
0.01 M HCl, kept on ice, should be used and added to tion is the major pathophysiological effect of alloxan,
test medium just prior to the start of the experiment in which results from the thiol group reactivity of alloxan.
order to obtain the final concentration. The reactive 5-carbonyl group of the alloxan molecule re-
± For injection stock solution should be diluted with ice- acts very avidly with thiol groups. The thiol groups of the
cold saline (0.9 % NaCl) immediately prior to injection. low affinity glucose phosphorylating enzyme glucoki-
± A toxic thiol reagent (Lenzen and Munday, 1991). nase, which functions in the beta cell as the glucose sen-
± It is reduced to dialuric acid in the presence of GSH sor for initiation of glucose-induced insulin secretion
and other thiols (Winterbourn and Munday, 1989). (Lenzen and Panten, 1988b), are particularly sensitive to
± During redox cycling between alloxan and dialuric oxidation by alloxan. With a half maximal inhibitory con-
acid, new alloxan is formed continuously, so that toxic centration in the range between 1 and 10 lmol/l the glu-
ROS species can be generated intracellularly over a cokinase protein is the most sensitive thiol enzyme
long time period (>1 h) (Lenzen and Munday, 1991). known in the beta cell (Lenzen and Panten, 1988a; Tiedge
± During each redox cycle a small amount of ªCom- et al., 2000).
pound 305º, an alloxan-GSH adduct of unknown struc- There is a great overall parallelism between the charac-
ture with a characteristic absorbance at 305 nm that is teristics of alloxan-induced inhibition of glucose-stimu-
not toxic, is formed (Patterson et al., 1949; Winter- lated insulin secretion and of alloxan inhibition of pan-
bourn and Munday, 1989; Lenzen and Munday, 1991; creatic beta cell glucokinase.
BroÈmme et al., 2002). i) Alloxan (Miwa et al., 1984; Hara et al., 1986; Me-
± A protoxin, which generates in its xenobiotic metabo- glasson et al., 1986; Lenzen et al., 1987b) inhibits gluco-
lism toxic ROS species (superoxide radicals (O·2 ±), hy- kinase because the reactive 5-carbonyl group reacts with
drogen peroxide (H2O2) and, in the presence of an iron thiols. The exceptional reactivity of the 5-carbonyl group
catalyst, hydroxyl radicals (·OH), when it redox cycles of alloxan is reflected in the facile reduction to the corre-
with dialuric acid (Munday, 1988; Winterbourn and sponding alcohol (dialuric acid) with reducing agents
Munday, 1989; Elsner et al., 2006). such as the thiols glutathione, cysteine and dithiothreitol
(Munday, 1988; Lenzen and Munday, 1991). Dialuric
acid, the reduction product of alloxan, is not thiol reactive
Biological effects of alloxan due to hydrophilicity, and therefore does not inhibit glucokinase as long as it is
chemical instability and glucose similarity of the kept in its reduced form through reducing agents (Lenzen
alloxan molecule: Selective uptake et al., 1988b; Lenzen and Munday, 1991). This underlines
and accumulation by the beta cell the importance of the free reactive 5-carbonyl group of
alloxan for its reaction with the thiol groups of the enzyme.
Alloxan is a hydrophilic and very instable chemical com- ii) The selectivity of the inhibition of insulin secretion
pound (Lenzen and Munday, 1991) and the alloxan mole- after exposure to alloxan (Ishibashi et al., 1978; Weaver
cule has a molecular shape, which is very similar to that et al., 1978b,c; Tait et al., 1983; Miwa et al., 1986) is re-
of glucose (Weaver et al., 1979; Gorus et al., 1982). Due stricted to that induced by glucose and its epimer man-
to their hydrophilic character both alloxan and glucose do nose, both of which induce insulin secretion through
not penetrate the lipid bilayer of the plasma membrane. binding to the glucose sensor glucokinase (Lenzen and
However, the molecular shape of alloxan is so similar to Panten, 1988b). This selective inhibition is due to the ex-
that of glucose, that the GLUT2 glucose transporter in the clusive sensitivity towards oxidation of the thiol groups of
beta cell plasma membrane accepts this molecule as a this enzyme by alloxan (Lenzen and Panten, 1988a). The
glucose analogue (glucomimetic) and transports it into inhibition of glucokinase reduces glucose oxidation and
the cytosol (Weaver et al., 1978a; Gorus et al., 1982). Al- ATP generation (Gunnarsson and HellerstroÈm, 1973;
loxan does not inhibit the function of the transporter with- Borg et al., 1979), thereby suppressing the signal generat-
in the biological lifetime of the chemical compound al- ing metabolic flux and the generation of the ATP signal
loxan (Elsner et al., 2002), so that the alloxan molecule for glucose-induced insulin secretion (Lenzen and Panten,Alloxan and Streptozotocin Diabetes 123 1988a). Inhibition of glucokinase is achieved within one calmodulin-dependent protein kinase (Colca et al., minute of exposure to alloxan and the result is a selective 1983), and aconitase (Lenzen and Mirzaie-Petri, 1992) inhibition of glucose-induced insulin secretion within as well as other proteins. minutes (Weaver et al., 1978c; Lenzen et al., 1988a). In- Thus, it can be concluded that the inhibition of glu- sulin biosynthesis is also inhibited by alloxan (Jain and cose-induced insulin secretion by alloxan is the result of Logothetopoulos, 1976; Niki et al., 1976), most likely the exquisite thiol reactivity of the glucose sensor gluco- through the same mechanism. kinase. ROS are not involved in this effect of alloxan. iii) The inhibition of glucose-induced insulin secretion is preceded by a very transient (1±2 min) stimulation of insulin secretion from the beta cells immediately after Biological effects of alloxan due to redox cycling and the exposure to alloxan (Weaver et al., 1978c). This effect generation of toxic reactive oxygen species (ROS): of alloxan can be explained by a reduction of ATP con- Beta cell toxicity and diabetogenicity sumption due to blockade of glucose phosphorylation by glucokinase (Lenzen and Panten, 1988b), resulting from Alloxan diabetes (Dunn and McLetchie, 1943; Goldner the instantaneous oxidation of catalytically important and Gomori, 1943; McLetchie, 1982) is a form of insu- thiol groups of this enzyme (Lenzen et al., 1988a; Tiedge lin-dependent diabetes mellitus caused by a direct toxic et al., 2000). This results in a transient increase of ATP in effect upon the endocrine pancreas; occlusion of the arter- the beta cell, which, through interaction with the ATP- ial blood supply to the pancreas prevents access of the in- sensitive potassium channel, causes this transient release jected alloxan to the beta cells and thus diabetes (Gomori, of insulin, before insulin secretion in response to glucose 1945; Bailey et al., 1950; Hellman and Diderholm, 1955). stimulation is halted permanently through suppression of Diabetes is the result of the selective pancreatic beta cell the mitochondrial metabolic flux and the ATP signal gen- toxicity of this compound which induces a necrosis of the erated in this process. beta cells (Dunn et al., 1943a,b; JoÈrns et al., 1997). iv) The insulin secretory response to other nutrient se- i) In order to destroy insulin-producing cells, alloxan cretagogues such as 2-ketoisocaproic acid and leucine or must enter the cell. Due to its hydrophilicity (Lenzen non-nutrient secretagogues such as the sulfonylurea drug and Munday, 1991) it does not permeate the lipid barrier tolbutamide remains initially intact, because it is not of the plasma membrane. However, due to its similarity mediated via glucokinase (Lenzen et al., 1988a). The se- with the glucose molecule, it can enter the cell via the cretory response to these secretagogues is lost only after a low affinity GLUT2 glucose transporter in the plasma delay of up to one hour after alloxan exposure (Borg, membrane (Elsner et al., 2002, 2003). This has been prov- 1981). This is a consequence of gradual deterioration of en by the fact that alloxan is not toxic to insulin-produc- beta cell function due to oxidation of functionally essen- ing cells which do not express the GLUT2 glucose trans- tial proteins as well as of the toxicity of free oxygen radi- porter (Bloch et al., 2000; Elsner et al., 2002). Alloxan cals generated during persisting continuous intracellular has no immediate and direct inhibitory effect upon glu- redox cycling between alloxan and dialuric acid. These cose transport (Elsner et al., 2002). toxic actions affect in particular enzymes of mitochon- ii) Since the half-life of alloxan is short (Patterson drial metabolism. et al., 1949; Lenzen and Munday, 1991), it must be taken v) Thiols such as GSH, cysteine and dithiotreitol pro- up and accumulated quickly in the beta cell (Hammar- tect glucokinase against alloxan inhibition because they stroÈm et al., 1967). After a few minutes, in aqueous solu- reduce alloxan to dialuric acid (Lenzen et al., 1988b; tion, alloxan has been converted into non-diabetogenic al- Lenzen and Mirzaie-Petri, 1991). By keeping dialuric loxanic acid through spontaneous decomposition (Lenzen acid reduced, a formation of the thiol reactive alloxan and Munday, 1991). Thus, alloxan is ineffective, when molecule is prevented (Lenzen and Munday, 1991). blood flow to the pancreas has been interrupted for a vi) However, only dithiols such as dithiotreitol but not few minutes after alloxan injection (Gomori, 1945; Bai- monothiols such as GSH or cysteine (Lenzen et al., ley et al., 1950; Hellman and Diderholm, 1955). 1988b; Lenzen and Mirzaie-Petri, 1991) can readily re- iii) N-substituted alloxan derivatives with a long car- verse alloxan-induced glucokinase inhibition. They bon side chain such as butylalloxan differ chemically in achieve this effect by reducing certain functionally essen- one respect from alloxan. In contrast to the hydrophilic tial cysteine residues of the glucokinase enzyme (Tiedge alloxan, they are lipophilic (Munday et al., 1993). Butyl- et al., 2000), which are oxidized through alloxan action alloxan (Fig. 1), like alloxan, is thiol reactive and gener- (Lenzen et al., 1988b; Lenzen and Mirzaie-Petri, 1991). ates ROS (Munday et al., 1993) and therefore can act in a At high concentrations, alloxan can inhibit many cellu- similar manner. In particular beta cells are damaged pre- lar functions. To this effect contributes significantly the ferentially, when isolated pancreatic islets are exposed to ability to oxidize thiol groups of many functionally im- lipophilic alloxan derivatives (BruÈckmann and Werthei- portant enzymes e. g. hexokinase (Lenzen et al., 1990), mer, 1947; JoÈrns et al., 1997). But as derivatives such as phosphofructokinase (Miwa et al., 1984), the Ca2+- and butylalloxan are lipophilic they can penetrate a plasma
124 Sigurd Lenzen
membrane without GLUT2 glucose transporter expres- v) Reduction of alloxan (A) to dialuric acid (AH2) in
sion (Elsner et al., 2002). Thus, these derivatives can en- the cell requires the presence of a suitable thiol, typically
ter all cell types and are systemically toxic rather then the tripeptide glutathione (GSH) to generate the redox cy-
diabetogenic (BruÈckmann and Wertheimer, 1947). Ne- cling partner (dialuric acid), while glutathione (GSSG) is
phrotoxicity is a dominating feature of the toxicity of li- oxidized (Reaction viii) (Winterbourn and Munday, 1989;
pophilic derivatives after systemic administration (BruÈck- BroÈmme et al., 2000).
mann and Wertheimer, 1947). This nephrotoxity, which A 2GSH ! AH2 GSSG viii
develops in rats after injection, is so severe that it causes
fatal renal failure in the animals before diabetes can de- The reaction takes place in two steps with the alloxan
velop (BruÈckmann and Wertheimer, 1947). Though at radical (·AH) as an intermediate product (Reactions ix±x).
the time of these early studies not known (BruÈckmann A GSH ! AH GS ix
and Wertheimer, 1947), the particular susceptibility of
the kidney to the toxic action of these lipophilic alloxan AH GSH ! AH2 GS x
derivatives can nowadays be explained as a result of a
preferential accumulation of these lipophilic toxins in vi) Other intracellular thiols, present at lower concen-
the tubular cells of the kidney, which, like the pancreatic trations in the cell, such as the monothiol cysteine and
beta cells, express the GLUT2 glucose transporter (Tho- other thiols and dithiols as well as ascorbic acid, are also
rens et al., 1988). Thus, experiments with lipophilic al- suitable reducing agents and thus may contribute, even
loxan derivatives have added much experimental support though to a lesser extent, to alloxan reduction (Elsner
to the concept that the hydrophilic character of the allox- et al., 2006). Reaction with thiols and thus generation of
an molecule is the decisive chemical property of this ROS is also possible with proteins such as enzymes (Len-
compound, which is responsible for selective accumula- zen et al., 1990; Lenzen and Mirzaie-Petri, 1991, 1992)
tion in the beta cells and thus for its diabetogenicity and albumin (Sakurai and Miura, 1989). During each re-
(Munday et al., 1993; JoÈrns et al., 1997). dox cycle a small amount of ªCompound 305º, an allox-
iv) Alloxan (Fig. 1) can generate ªreactive oxygen spe- an-GSH adduct (Patterson et al., 1949; Winterbourn and
ciesº (ROS) in a cyclic reaction between this substance Munday, 1989; Lenzen and Munday, 1991; BroÈmme
and its reduction product, dialuric acid (Cohen and Heik- et al., 2002), is formed through reaction between alloxan
kila, 1974; Munday, 1988; Winterbourn et al., 1989; Win- and GSH. While the intracellular concentration of ªCom-
terbourn and Munday, 1989) (Reactions i±ii). The beta pound 305º increases in a time-dependent manner, the
cell toxic action of alloxan (A) is initiated by free radicals amount of reduced GSH available in the cell for redox cy-
formed in this redox reaction (Cohen and Heikkila, 1974; cling diminishes gradually and thus fosters a lower pro-
Munday, 1988; Oberley, 1988; Winterbourn et al., 1989; oxidative ratio between alloxan and GSH, rather than a
Winterbourn and Munday, 1989). Autoxidation of dia- higher antioxidative ratio (Munday, 1988; Winterbourn
luric acid (AH2) generates superoxide radicals (O·2 ±) (Re- et al., 1989; Winterbourn and Munday, 1989).
actions iii±iv) and hydrogen peroxide (H2O2) (Reactions vii) Thiols, cysteine as well as GSH, have been re-
iii±iv) and, in the Fenton reaction (Reaction v), in the pre- ported long ago already to protect rats against develop-
sence of a suitable metal catalyst (typically iron) (Reac- ment of alloxan diabetes, when injected together with al-
tion vi), hydroxyl radicals (·OH) (Reactions v±vii). The loxan (Lazarow, 1946; Lazarow et al., 1948; Sen and
autoxidation of dialuric acid involves the intermediate Bhattacharya, 1952). This, at a first glance, paradoxical
formation of the alloxan radical (·AH) (Reactions i±iv) observation can now be explained in the light of the un-
(Munday, 1988; Winterbourn et al., 1989; Winterbourn derstanding of the molecular mechanism of alloxan ac-
and Munday, 1989). tion. When concentrations of reducing agents in the blood
stream or in the extracellular space are significantly in-
AH2 O2 ! AH O2 H i creased through injection of a thiol, more alloxan is re-
duced extracellularly so that less is available for accumu-
AH O2 ! A O2 H ii lation in the beta cells, thereby ameliorating the beta cell
toxic and diabetogenic action of alloxan. Normally the
AH2 O2 H ! AH H2 O2 iii capacity for reduction of alloxan, redox cycling and gen-
eration of ROS in the circulation (Sakurai and Miura,
AH O2 H ! A H2 O2 iv 1989) is not sufficient to prevent the alloxan molecule to
reach and enter the pancreatic beta cell. On the other
H2 O2 e ! HO HO v hand, through fostering of redox cycling in the organism,
the general systemic toxicity of alloxan is increased.
FeII H2 O2 ! FeIII OH OH vi viii) The major oxidation pathway of dialuric acid, an
metal
O·2 ±-dependent chain reaction, is inhibited by superoxide
Net : O2 H2 O2 ! O2 OH OH vii dismutase (SOD) (Reaction xi). In the presence of SOD,
catalystAlloxan and Streptozotocin Diabetes 125
an autocatalytic process involving the interaction between suppression of dialuric acid autoxidation prevents ROS
dialuric acid and alloxan becomes important (Munday, generation (Munday, 1988; Winterbourn and Munday,
1988), while in the presence of a transition metal, a third 1989) and through this mechanism counteracts the toxi-
oxidation mechanism, dependent upon H2O2, has been city to insulin-producing cells (Elsner et al., 2006). In-
identified (Munday, 1988). This latter step is inhibited creasing the concentrations of the toxins, however, can re-
by the hydrogen peroxide inactivating enzyme catalase install the toxic action of both compounds. This is due to
(Munday, 1988) (Reaction xii). The other hydrogen per- the fact that an autocatalytic reaction between the oxi-
oxide inactivating enzyme, glutathione peroxidase dized and reduced pyrimidine comes into action, which
(GPx), can principally act in a similar manner. But this also generates ROS, when the chain reaction is sup-
enzyme requires GSH, which is oxidized in this reaction pressed by SOD (Munday, 1988; Munday and Winter-
to GSSG (Reaction xiii). bourn, 1989; Winterbourn and Munday, 1989).
xiii) The superoxide radical, however, is apparently not
2H 2O2 ! H2 O2 O2 xi
SOD the species responsible for the cytotoxicity of alloxan and
dialuric acid, because the hydrogen peroxide (H2O2) inac-
2H2 O2 ! O2 2H2 O xii tivating enzyme catalase provided significantly better
Catalase
protection against the toxicity to insulin-producing cells
GPx than the superoxide radical (O·2 ±) inactivating enzyme
2GSH H2 O2 ! 2H2 O GSSG xiii
SOD, though catalase does not prevent redox cycling
ix) The alloxan molecule itself is not cytotoxic. When and therefore does not prevent superoxide radical forma-
kept in the oxidized form and when reduction and/or re- tion (Munday, 1988; Winterbourn and Munday, 1989).
dox cycling are prevented, alloxan does not generate ROS This proves convincingly that not the superoxide radical
(Elsner et al., 2006). Without ROS generation alloxan is (O·2 ±) but rather the hydroxyl radical (·OH) is the ultimate
not toxic to insulin-producing cells (Elsner et al., 2006). toxic ROS species, whose formation from hydrogen per-
x) The reduction product dialuric acid itself is also not oxide (H2O2) is prevented through destruction of hydro-
toxic, when kept in the reduced form and when oxidation gen peroxide by catalase (Munday, 1988; Winterbourn
and redox cycling and/or generation of ROS is prevented and Munday, 1989).
(Munday, 1988; Winterbourn and Munday, 1989; Elsner xiv) Optimal protection against the cytotoxic action of
et al., 2006). SOD and catalase prevent dialuric acid toxi- alloxan and dialuric acid towards insulin-producing cells,
city as documented by the protection against death of in- however, is provided only by a combination of SOD plus
sulin-producing cells (Elsner et al., 2006). catalase. Only this combination completely prevents re-
xi) At variance from alloxan, dialuric acid autooxidizes dox cycling between alloxan and dialuric acid and thus
spontaneously in the presence of O2, thus generating cy- the generation of all ROS species in this chain reaction,
totoxic ROS in the absence of a thiol (Munday, 1988; namely superoxide radicals (O·2 ±), hydrogen peroxide
Winterbourn et al., 1989), even when not taken up into (H2O2) and hydroxyl radicals (·OH), which are ultimately
the cell. In contrast, alloxan, when restricted to the extra- responsible for cell death (Munday, 1988; Winterbourn
cellular space, is not toxic in the absence of thiols such as and Munday, 1989).
GSH (Elsner et al., 2006). Thus it is not surprising that xv) A multitude of metal and in particular specific iron
dialuric acid can also induce diabetes (BruÈckmann and chelators as well as hydroxyl scavengers have been tested
Wertheimer, 1945, 1947; Bailey et al., 1946; Saviano and for their protective action both in vitro and in vivo. But
De Franciscis, 1946; Merlini, 1951) and causes the same though metal chelators such as EDTA, diethylenetriamine
beta cell lesions as alloxan (JoÈrns et al., 1997). Thiols in pentaacetic acid, desferrioxamine and phenanthroline and
the plasma membrane, with which alloxan could interact ROS scavengers such as the alcohols butanol and di-
and in the consequence could be reduced and generate methylsulfoxide and the urea derivatives dimethylurea
ROS in a redox cycle, are apparently not present or not and thiourea have been shown in many experimental sit-
accessible to an extent, which would allow generation of uations to provide protection against alloxan toxicity, the
ROS sufficient to damage the cells (Elsner et al., 2006). results obtained have never been entirely unequivocal
Thus, a former hypothesis that alloxan might be cytotoxic (Heikkila et al., 1974, 1976; Heikkila, 1977; Heikkila
due to interaction with thiol groups in the beta cell mem- and Cabbat, 1978, 1982; Grankvist et al., 1979a; Tibaldi
brane (Watkins et al., 1970) is not valid. et al., 1979; Fischer and Hamburger, 1980a,b; Fischer
xii) The antioxidative enzyme SOD greatly attenuates and Harman, 1982; JoÈrns et al., 1999). Due to the ex-
the toxicity of both alloxan and dialuric acid to insulin- tremely short half-life and the extraordinary chemical re-
producing cells in the presence of GSH. The reason for activity of the hydroxyl radical, it is not surprising that a
this cytoprotective effect is the ability of SOD to scav- complete protection against alloxan toxicity cannot be
enge superoxide radicals, which are generated in the achieved, because hydroxyl radicals interact with biologi-
O2± dependent chain reaction between dialuric acid and cal targets before they can be inactivated through hydrox-
alloxan (Munday and Winterbourn, 1989). The resultant yl scavengers. In the same way, it is not surprising that it126 Sigurd Lenzen
has been extremely difficult to efficiently suppress metal Munday, 1989). Not surprisingly, therefore, effective pre-
catalysed hydroxyl radical formation through metal che- vention of redox cycling and generation of ROS or effi-
lators (Grankvist et al., 1979b; Fischer and Hamburger, cient inactivation of ROS can prevent beta cell death
1980b; Fischer and Harman, 1982; Heikkila and Cabbat, (JoÈrns et al., 1999; Elsner et al., 2006) and counteract the
1982; JoÈrns et al., 1999), since it is crucial that metal che- development of alloxan diabetes in vivo (Heikkila et al.,
lation is achieved before hydroxyl radical formation is in- 1976).
itiated. Due to the chemical properties of the chelators xviii) Thus, it can be concluded that the pancreatic beta
and scavengers, this cannot always be achieved comple- cell toxicity and the resultant diabetogenicity of alloxan is
tely, unless experimental conditions are optimized. This due to the redox cycling and the generation of toxic reac-
conclusion is supported by the observation that the toxi- tive oxygen species (ROS) in combination with the hy-
city of alloxan and dialuric acid to insulin-producing cells drophilicity and the glucose similarity of the molecular
in vitro is suppressed through the iron chelator desferriox- shape of alloxan.
amine (Elsner et al., 2006), which prevents the generation
of the very toxic hydroxyl radicals in the iron catalyzed
Fenton reaction (Munday, 1988; Winterbourn and Mun- Streptozotocin mechanism of action
day, 1989). Overall, the results with radical scavengers
and metal chelators support the contention, that the hy- The diabetogenic agent streptozotocin (STZ) (Fig. 3) in-
droxyl radical is the ROS species ultimately responsible hibits insulin secretion and causes a state of insulin-de-
for alloxan-induced cell death. pendent diabetes mellitus through its ability to induce a
xvi) The intracellular GSH concentrations are in the selective necrosis of the pancreatic beta cells. Both ef-
same millimolar concentration range in insulin-producing fects can be assorted to the alkylating potency of strepto-
cells as in liver cells. As both cell types express the zotocin. The common chemical denominator of the two
GLUT2 glucose transporter a difference in the intracellu- effects is the selective cellular uptake and the accumula-
lar GSH concentration cannot explain the much greater tion of streptozotocin by the beta cell.
susceptibility of insulin-producing cells than of liver cells
to the toxicity of alloxan in vivo (Rerup, 1970). However,
liver cells are much better endowed with the hydrogen
peroxide inactivating enzyme catalase than insulin-
producing cells (Grankvist et al., 1979a; Lenzen et al.,
1996a; Tiedge et al., 1997). When the low intracellular le-
vels of catalase protein expression are raised in insulin-
producing cells through overexpression, these cells are
Fig. 3: Chemical formulas of streptozotocin (STZ) and methyl-
protected equally well (Elsner et al., 2006). This proves
nitrosourea (MNU).
convincingly that the low hydrogen peroxide (H2O2) in-
activating enzyme capacity is crucially responsible for
the exquisite susceptibility of insulin-producing beta cells Chemical properties of streptozotocin
towards alloxan toxicity.
xvii) Thus the mechanism underlying the cytotoxic ac- Streptozotocin (2-deoxy-2-(((methylnitrosoamino)carbo-
tion of alloxan to insulin-producing cells is due to reduc- nyl)amino)-D-glucopyranose) is:
tion by interaction with intracellular thiols such as GSH ± A cytotoxic methylnitrosourea moiety (N-methyl-N-ni-
and the resultant formation of cytotoxic ROS in a cyclic trosourea) attached to the glucose (2-deoxyglucose)
reaction between alloxan and its reduction product, dia- molecule. It is also a glucosamine derivative.
luric acid (Munday, 1988; Winterbourn and Munday, ± A beta cell toxic glucose analogue.
1989). The autoxidation of the latter generates superoxide ± A hydrophilic compound.
radicals (O·2 ±), and hydrogen peroxide (H2O2) (Munday, ± An alkylating agent (Bennett and Pegg, 1981).
1988; Winterbourn and Munday, 1989) and ultimately, ± Relatively stable at pH 7.4 and 37 °C (at least for up to
from the latter, the hydroxyl radical (·OH) (Munday, one hour) (Axler, 1982; Lee et al., 1993; Povoski et al.,
1988; Winterbourn and Munday, 1989). Alloxan diabetes, 1993).
induced in laboratory animals through injection of this ± For in vitro experimentation concentrated stock solu-
diabetogenic compound, is therefore the result of selec- tions in 0.01 M HCl, kept on ice, should be used and
tive uptake of alloxan via the low affinity GLUT2 glucose added to test medium just prior to the start of the ex-
transporter (Elsner et al., 2002) into a pancreatic beta cell, periment in order to obtain the final concentration.
which due to its weak expression of hydrogen peroxide ± For injection a stable solution in citrate buffer (pH 4.5)
inactivating enzymes (Lenzen et al., 1996a; Tiedge et al., is most suited.
1997) is badly protected against hydroxyl radical (·OH)
mediated cytotoxicity (Munday, 1988; Winterbourn andAlloxan and Streptozotocin Diabetes 127
Biological effects of streptozotocin due to this transporter such as hepatocytes and renal tubular cells
hydrophilicity and glucose similarity (Thorens et al., 1988). Therefore any streptozotocin treat-
of the streptozotocin molecule: Beta cell selectivity ment of animals leads not only to diabetes but can also
of biological effects cause liver and kidney damage of variable degree (Rerup,
1970). A number of chemically very different alkylating
The toxic action of streptozotocin and chemically related agents, which have been studied for their beta cell toxi-
alkylating compounds requires their uptake into the cells. city (Delaney et al., 1995), are not dependent for their
Typically nitrosoureas are lipophilic and tissue uptake is toxic action upon the expression of the GLUT2 glucose
quick. However, due to the hexose substitution streptozo- transporter (Elsner et al., 2000). Therefore it is not sur-
tocin is less lipophilic and this significantly affects the prising that these compounds are not selectively toxic to
cellular distribution. The greater hydrophilicity of strepto- beta cells and not diabetogenic (Delaney et al., 1995; Els-
zotocin reduces its chance to enter cells without support ner et al., 2000).
and also prevents access to the brain via the blood-brain
barrier. Streptozotocin, in contrast to other nitrosoureas
including N-methyl-N-nitrosourea (MNU) (Fig. 3), causes Biological effects of streptozotocin due to alkylation:
little myelosuppression or carbamoylation of lysine resi- Inhibition of beta cell function and insulin secretion
dues of proteins (Bennett and Pegg, 1981; Weiss, 1982).
On the other hand, through the attachment of the methyl- The effects of streptozotocin on glucose and insulin
nitrosourea moiety to the 2 carbon of glucose as a carrier homeostasis reflect toxin-induced abnormalities in pan-
molecule streptozotocin is selectively accumulated in creatic beta cell function. Initially an inhibition of insulin
pancreatic beta cells (Karunanayake et al., 1976; TjaÈlve biosynthesis and glucose-induced secretion as well as an
et al., 1976) and becomes a beta cell toxic and diabeto- impairment of glucose metabolism, both glucose oxida-
genic compound, though it also exhibits significant renal tion and oxygen consumption (Strandell et al., 1988;
and hepatic toxicity (Rerup, 1970; Weiss, 1982). Nukatsuka et al., 1990; Bedoya et al., 1996), become evi-
Streptozotocin has also been employed in the treatment dent. On the other hand, streptozotocin has no immediate
of human islet-cell carcinomas and malignant carcinoid and direct inhibitory effect upon glucose transport (Elsner
tumours (Schein et al., 1974). However, clinical strepto- et al., 2000) or upon glucose phosphorylation through
zotocin cancer treatment typically involves repetitive, glucokinase (Lenzen et al., 1987b). However, at later stages
miniscule doses significantly lower than necessary to in- of functional beta cell impairment, deficiencies at the level
duce diabetes in experimental animals. This together with of gene and protein expression and function of these struc-
the resistance of human beta cells to streptozotocin toxi- tures become apparent (Wang and Gleichmann, 1998).
city explains why diabetes is not a typical side effect of Even before negative effects due to mitochondrial
streptozotocin treatment though renal and hepatic toxicity DNA and protein alkylation become evident, a streptozo-
are often encountered in these patients (Schein et al., tocin-induced depletion of NAD+ may result in an inhibi-
1974; Weiss, 1982). tion of insulin biosynthesis and secretion (Yamamoto
The selective pancreatic beta cell toxicity of streptozo- et al., 1981a,b; Uchigata et al., 1982; Strandell et al.,
tocin and the resulting diabetic metabolic state are clearly 1988). Later, inhibition of glucose-induced and also of
related to the glucose moiety in its chemical structure, amino acid-induced (Eizirik et al., 1988a,b) insulin secre-
which enables streptozotocin to enter the beta cell via tion, due to dysfunction of mitochondrial enzymes
the low affinity GLUT2 glucose transporter in the plasma (Rasschaert et al., 1992) and damage to the mitochondrial
membrane (Elsner et al., 2000). This hypothesis is sup- genome (Eizirik et al., 1991) become apparent. This im-
ported by the observation that insulin-producing cells that pairment is more marked for nutrient than for non-nutri-
do not express this glucose transporter, are resistant to ent induced insulin secretion. This interpretation has been
streptozotocin toxicity (Ledoux and Wilson, 1984; Elsner confirmed through studies which have shown that pre-
et al., 2000) and become sensitive to the toxic action of treatment of isolated pancreatic islets with the poly
this compound only after expression of the GLUT2 glu- (ADP-ribose) polymerase (PARP) inhibitor nicotinamide
cose transporter protein in the plasma membrane prevents early inhibition of beta cell function during the
(Schnedl et al., 1994; Elsner et al., 2000). This observa- first day after streptozotocin exposure, while long-term
tion explains the greater toxicity of streptozotocin, when inhibition of insulin secretion six days after streptozoto-
compared with N-methyl-N-nitrosourea (MNU) in cin exposure was not counteracted by nicotinamide
GLUT2 expressing beta cells even though both sub- (Strandell et al., 1989).
stances alkylate DNA to a similar extent (Ledoux et al., When exposed to high cytotoxic streptozotocin con-
1986; Wilson et al., 1988; Elsner et al., 2000). The impor- centrations, these initial circumscribed functional defects
tance of the GLUT2 glucose transporter for the toxic ac- gradually turn into more severe functional deficiencies.
tion of streptozotocin is also confirmed by the observa- These changes are of a more general and unspecific nat-
tion of streptozotocin damage to other cells expressing ure and result from the progressive deterioration of the128 Sigurd Lenzen
function of a dying beta cell. Of note, at lower streptozo- NAD+ and the subsequent loss of ATP (Burkart et al.,
tocin concentrations, many beta cells are able to survive 1999; Masutani et al., 1999; Pieper et al., 1999a,b) and
the initial insult, but keep a long-term functional defect thus cell death.
characterized by a preferential deficiency of mitochon- The role of alkylation in beta cell damage has also
drial oxidative metabolism (Eizirik et al., 1988a,b, 1991; been examined by the use of the ethylating agents N-
Rasschaert et al., 1992). ethyl-N-nitrosourea and ethyl methanesulphonate. These
ethylating agents are known to be less toxic than their
methylating counterparts (Karran and Bignami, 1992; De-
Biological effects of streptozotocin due to alkylation: laney et al., 1995). This has been attributed to O6-ethyl-
Beta cell toxicity and diabetogenicity guanine being less toxic than O6-methylguanine (Karran
and Bignami, 1992). The lower toxicity of the ethylating
Toxicity of streptozotocin and related compounds resides agents has been interpreted as evidence for a role of O6-
in their ability to alkylate biological macromolecules alkylguanine in the mechanism of the toxic action of this
(Bennett and Pegg, 1981). It is generally assumed that group of alkylating compounds. The fact that N-ethyl-N-
the toxic activity of streptozotocin relates to the DNA al- nitrosourea and ethyl methanesulphonate are significantly
kylating activity of its methylnitrosourea moiety (Bennett less toxic to insulin-producing cells than MNU and
and Pegg, 1981; Uchigata et al., 1982; Ledoux et al., methyl methanesulphonate (Delaney et al., 1995; Elsner
1986; Wilson et al., 1988; Murata et al., 1999), especially et al., 2000) has been taken as support for the contention
at the O6 position of guanine (Goldmacher et al., 1986; that in insulin-producing cells, like in other cell types, the
Green et al., 1989; Karran and Bignami, 1992). This mechanism of toxic action is due to alkylation (Bennett
DNA damage results, along a defined chain of events and Pegg, 1981), with the methylation of DNA bases
(Pieper et al., 1999b), in a fragmentation of the DNA (Ya- being more toxic than the ethylation (Karran and Bigna-
mamoto et al., 1981a,b; Morgan et al., 1994) and ulti- mi, 1992).
mately in the necrosis of the pancreatic beta cells (Lenzen An alternative hypothesis proposes that part of the dia-
et al., 1996b) through depletion of cellular energy stores. betogenic effect of streptozotocin may relate not to its al-
It is the resultant activation of PARP that, in an attempt to kylating ability but to its potential to act as an intracellu-
repair the damaged DNA, depletes the cellular NAD+ and lar nitric oxide (NO) donor (Turk et al., 1993). Both
subsequently the ATP stores (Schein and Loftus, 1968; streptozotocin and MNU contain a nitroso group (Fig. 3)
Yamamoto et al., 1981a,b; Uchigata et al., 1982; Sandler and can liberate nitric oxide (Rogers and Ignarro, 1992;
and Swenne, 1983) due to overstimulation of the DNA re- Kwon et al., 1994; Kaneto et al., 1995; KroÈncke et al.,
pair mechanisms (Pieper et al., 1999b). Though strepto- 1995; Tanaka et al., 1995; Bedoya et al., 1996; Murata
zotocin also methylates proteins (Bennett and Pegg, et al., 1999) similar to that of other nitric oxide donors
1981; Wilson et al., 1988) DNA methylation is ultimately such as sodium nitroprusside or 3-morpholinosydnoni-
responsible for beta cell death. But it can be anticipated mine (Tiedge et al., 1999). In fact streptozotocin has been
that the methylation of beta cell proteins also contributes shown to increase the activity of guanylyl cyclase and the
to the functional defects of the beta cells after exposition formation of cGMP, which are characteristic effects of
to streptozotocin in vitro or in vivo. NO. However, the alkylating agent methyl methanesul-
Inhibitors of poly ADP-ribosylation suppress this pro- phonate is the most toxic compound though, unlike
cess of DNA methylation. Thus, injection of nicotinamide MNU, it is not a nitric oxide donor (Delaney et al.,
and other poly (ADP-ribose) polymerase inhibitors in 1995), indicating that NO is not an indispensable prere-
parallel with or prior to the administration of streptozoto- quisite for the toxic action of this group of alkylating
cin is well known to protect beta cells against the toxic agents including the diabetogenic compound streptozoto-
action of streptozotocin and to prevent the development cin. This view is convincingly supported by the observa-
of a diabetic state (Schein et al., 1967; Dulin and Wyse, tion that mice deficient in PARP are resistant to beta cell
1969; Stauffacher et al., 1970), however, on account of death mediated by streptozotocin in spite of DNA frag-
the later development of insulin-producing pancreatic tu- mentation. The absence of PARP, which is activated by
mours (Rakieten et al., 1971; Preston, 1985; Lenzen DNA damage, prevents the depletion of the cofactor
et al., 1987a). Poly (ADP-ribose) polymerase inhibitors NAD+ and the subsequent loss of ATP (Burkart et al.,
protect islets also in vitro against streptozotocin-induced 1999; Masutani et al., 1999; Pieper et al., 1999a,b). Thus
inhibition of insulin biosynthesis and secretion through NO and free nitrous radicals (i. e. peroxinitrite) may be an
the same mechanism of action (Bedoya et al., 1996). This aggravating factor in the toxic action of streptozotocin but
concept has been convincingly confirmed by the observa- NO is apparently not crucial for its beta cell toxic action
tion that mice deficient in PARP are resistant to beta cell (Delaney et al., 1995; Yamamoto et al., 1981a,b; Elsner
death mediated by streptozotocin in spite of DNA frag- et al., 2000).
mentation. The absence of PARP, which is activated by An involvement of reactive oxygen species in the toxic
DNA damage, prevents the depletion of the cofactor action of streptozotocin, which may be produced duringAlloxan and Streptozotocin Diabetes 129
uric acid generation as the final product of ATP degrada- cell (De Vos et al., 1995; Ferrer et al., 1995) rather than
tion by xanthine oxidase from hypoxanthine, has also the inability of the human GLUT2 glucose transporter
been considered (Nukatsuka et al., 1990). And indeed, isoform to provide uptake of alloxan and streptozotocin
some indirect evidence for a participation of ROS has into the intracellular compartment which is responsible
been obtained in experiments with scavengers (Sandler for the extraordinarily high resistance of humans against
and Swenne, 1983; Okamoto, 1996). So some minor gen- the diabetogenic action of alloxan and streptozotocin.
eration of ROS including superoxide and hydroxyl radi- However, even if the human GLUT2 glucose transporter
cals originating from hydrogen peroxide dismutation dur- isoform would be more abundant in human pancreatic
ing hypoxanthine metabolism may accompany the effect beta cells the lower capacity for uptake of the toxins
of streptozotocin and accelerate the process of beta cell through the human GLUT2 glucose transporter isoform
destruction. However, it is the alkylating potency of strep- as compared to the rat transporter isoform would limit
tozotocin, which causes through ATP depletion a state of the diabetogenicity of both alloxan and streptozotocin in
energy deficiency similar to that, which results from other humans (Elsner et al., 2003). Therefore, it is also not sur-
states of hypoxia and ischemia and is thus crucial for its prising that development of diabetes is not known to be a
beta cell toxicity. typical side effect when streptozotocin is used as a che-
Alloxan causes not only protein but also DNA damage motherapeutic agent in human cancer treatment and that
(Takasu et al., 1991; Okamoto, 1996) through ROS toxi- it is not a particularly efficient anticancer drug when used
city. In the case of alloxan toxicity nicotinamide may pro- in the treatment of human insulinomas which usually do
vide protection through free radical scavenging (Ledoux not express the GLUT2 glucose transporter (Yang and
et al., 1988). Such a capacity for free radical scavenging Wright, 2002).
may also be a minor complementary but not crucial com- On the other hand there is no evidence for any differ-
ponent in the protective action of nicotinamide against al- ence in antioxidative equipment between rodent and hu-
loxan toxicity. man beta cells that would be sufficiently large (Welsh
et al., 1995) to explain these observed species differences
in susceptibility towards the toxicity of the two agents
Species differences of alloxan (Elsner et al., 2003). But, as shown in studies comparing
and streptozotocin toxicity different strains differences in antioxidative equipment
may contribute to differences in sensitivity to alloxan
Alloxan and streptozotocin induce insulin deficiency due toxicity in the same species (Mathews and Leiter,
to their selective pancreatic beta cell toxicity, but there 1999a,b).
are significant species differences in the diabetogenicity The exclusion of the toxins from the intracellular space
of alloxan and streptozotocin (Rerup, 1970). Rodents are due to a lack of GLUT2 glucose transporter protein ex-
particularly prone to the diabetogenic action of these two pression in the plasma membrane can provide a common
diabetogens, while humans are resistant to the toxicity explanation for the species differences in toxicity of both
(Rerup, 1970). Isolated human pancreatic beta cells, in compounds. A unifying concept on the basis of a differ-
contrast to rodent beta cells, are resistant to the toxic ac- ence in the intracellular milieu, on the other hand, would,
tion of alloxan and streptozotocin in vitro (Eizirik et al., in view of the different mechanisms of cytotoxic action of
1994; Tyrberg et al., 2001; Yang and Wright, 2002). Sev- alloxan and streptozotocin, require a separate explanation
eral studies have shown that after transplantation of hu- for each compound for these species differences in beta
man pancreatic islets into nude mice, beta cells were not cell toxicity and diabetogenicity. However, there is no
damaged even after injection of high doses of streptozo- convincing experimental basis for such an alternative uni-
tocin (Yang and Wright, 2002) or alloxan (Eizirik et al., fying concept. It is thus the exclusion of the toxins from
1994; Tyrberg et al., 2001), while the beta cells of rat is- the intracellular space due to a lack of GLUT2 glucose
lets transplanted concomitantly into the same animal transporter expression, which provides the common ex-
were destroyed (Eizirik et al., 1994; Tyrberg et al., 2001; planation for the resistance of human insulin-producing
Yang and Wright, 2002). cells against toxicity of these two beta cell toxins.
Rat insulin-producing tissue culture cells, which do not Rodents but also other species such as rabbits and dogs
constitutively express the GLUT2 glucose transporter, are are sensitive to the diabetogenic action of alloxan and
also resistant while expression of the rat GLUT2 glucose streptozotocin, even though there are significant species
transporter isoform renders such cells sensitive to the differences, so that different doses of the toxins are re-
toxicity of both toxins (Elsner et al., 2000, 2002). In a quired in different animal species in order to obtain the
study comparing insulin-producing cells, in which either same result (Rerup, 1970). The guinea pig is an example
the rat or the human GLUT2 glucose transporter isoform of a particularly insensitive species when the toxins are
had been expressed (Elsner et al., 2003), it could be prov- administered systemically. Injection of alloxan into the
en that it is apparently the very low level of constitutive circulation of the pancreas has been shown to cause beta
GLUT2 glucose transporter expression in the human beta cell necrosis and diabetes in the guinea, while systemicYou can also read
