Pathophysiology of Feline Diabetes Mellitus

7. Pathophysiology of Feline Diabetes Mellitus

Insulin Resistance in Feline Diabetes Mellitus

One of the two major metabolic hallmarks of human 2DM and feline DM, next to disturbed pancreatic beta-cell function, is insulin resistance. Insulin resistance, or lower than normal insulin sensitivity, is characterized by a reduced response of insulin target tissues to a given amount of insulin. This can be assessed via insulin-sensitive glucose uptake which is markedly reduced in insulin resistant individuals. While oversecretion of insulin may compensate at least partly for insulin resistance, measurable glucose intolerance or overt hyperglycemia will develop once hyperinsulinemia cannot be sustained, or when maintained stress on beta-cells leads to their exhaustion (Figure 8).

Figure 8. The vicious circle of insulin resistance, defect in beta-cell function and glucotoxicity, that eventually leads to beta-cell exhaustion and overt DM.

Tests to Assess Insulin Sensitivity

The classical clinical tests to assess insulin sensitivity and secretion are the intravenous glucose tolerance test (IVGTT; O’Brien et al., 1985; Appleton et al., 2001a, 2001b) or the insulin sensitivity test (IST; Feldhahn et al., 1999; Appleton et al., 2001a, 2001b). In the IVGTT, the increase in blood glucose and insulin concentrations are measured following an intravenous glucose bolus. Reported upper limits of the normal range for glucose half-life in plasma (glucose T1/2) in healthy cats are approximately 75 - 80 min (Lutz and Rand, 1996; Appleton et al., 2001a, 2001b). In the IST, the glucoselowering effect of insulin is assessed directly (Appleton et al., 2001a, 2001b).

Glucose intolerant "pre-diabetic" and diabetic cats typically present with higher glucose concentrations in IVGTTs and with glucose T1/2 that is prolonged. Fasting insulin levels seem to be more variable because they have been reported to be elevated in some studies (e.g., Nelson et al., 1990) but not in others (e.g., Lutz & Rand, 1996).

Mechanisms For Insulin Resistance

Impaired glucose tolerance in diabetic cats is the result of a reduced insulin response (O’Brien et al., 1985) and reduced insulin sensitivity. Insulin sensitivity in diabetic cats is approximately 6 times lower than in healthy cats (Feldhahn et al., 1999). The exact underlying mechanisms for insulin resistance in human 2DM and in feline DM are still unknown (Reaven, 2005; [Reusch et al., 2006b].). Similar to humans, the major cause of insulin resistance in cats is obesity and physical inactivity. Insulin sensitivity in obese cats is markedly reduced compared to lean control animals (see below).

Factors Contributing to Insulin Resistance

Genetic causes of receptor or post-receptor defects have not been analyzed in detail in cats, but some molecular tools have become available lately that will allow us to study some of the underlying mechanisms of peripheral insulin resistance in more detail. Most attention has been drawn to glucose transporters in insulin-sensitive tissues and to metabolically active cytokines released from adipose tissue (e.g., Brennan et al., 2004; Hoenig et al., 2007a; Zini et al., 2006).

Whether there is a systemic difference in insulin sensitivity between male and female cats is less clear. On the one hand, it has been reported that male cats have lower insulin sensitivity and higher baseline insulin concentrations than female cats (Appleton et al., 2001a; Rand & Marshall, 2005). The latter study was performed in lean animals which were fed a diet relatively high in carbohydrate. However, all animals, males and females, were castrated at the time of study. Therefore, it is unlikely that direct effects of sexual hormones can explain the difference in insulin sensitivity. Either early effects of sexual hormones, acting before the time of castration, or indirect effects of sexual hormones may account for these differences.

More research is needed to investigate the possible gender differences in insulin sensitivity and the development of feline DM. (© Yves Lanceau/RC (Chartreux)).

On the other hand, obesity is well recognized as the main risk factor to induce insulin resistance, and relative body weight (BW) gain after castration appears to occur more rapidly in females than in males (Martin & Siliart, 2005). This somehow contrasts to a study by Hoenig et al. (2007b) who reported that insulin leads to increased glucose oxidation in obese castrated males while castrated females maintain greater fat oxidation in response to insulin. This metabolic gender difference was therefore supposed to favor more rapid fat accumulation in males than females, which may explain the greater risk of DM in neutered males. However, the same authors also reported that gender was not an independent risk factor in a study comparing glucose kinetics parameters between lean and obese cats (Hoenig et al., 2007a, 2007b).

Other causes of insulin resistance include insulin antagonistic hormones, e.g., glucocorticosteroids and progestins, which directly counteract insulin action. Further, at least in other species, glucocorticosteroids increase food intake and may therefore contribute to the development of obesity. Presumably, they have similar effects in cats. Hyperthyroidism and growth hormone excess (acromegaly) have also been shown to reduce glucose tolerance, possibly due to the induction of peripheral insulin resistance (Hoenig & Ferguson, 1989; Feldman & Nelson, 2004).

Disturbed Pancreatic Beta-cell Function

The second major hallmark of feline diabetes is disturbed beta-cell function. Typical defects are a markedly reduced or missing first phase insulin secretion and a delayed onset of second phase insulin release which mainly relies on insulin synthesis. Even though the baseline insulin concentration may be unchanged, the overall insulin secretory capacity is clearly reduced in diabetic cats (Figure 9). In most cases, the underlying defect of disturbed beta-cell function at the molecular level is completely unknown.

Figure 9. Plasma amylin and plasma insulin concentrations in cats with normal and disturbed glucose tolerance.

Because insulin and amylin are usually cosecreted, similar defects also refer to amylin secretion (Figure 9). However, early phases of feline DM seem to be associated with relative hyperamylinemia (Lutz & Rand, 1996). It is currently unknown whether initial hypersecretion of amylin contributes to accelerated deposition of pancreatic islet amyloid (see below) or whether it may rather be regarded as an adaptive response to help control blood glucose due to amylin’s metabolic effects such as inhibition of postprandial glucagon secretion (see below).

Once established, deficient insulin secretion leads to overt hyperglycemia. Sustained hyperglycemia then causes progressive disruption of normal betacell function. This phenomenon is called glucotoxicity (Prentki et al., 2002) and will be discussed below. Further complication results from inflammatory events which are now considered an important feature in the pathophysiological sequence leading to beta-cell insufficiency in 2DM like syndromes (Donath et al., 2005; see below).

Obesity and the Development of Diabetes Mellitus

The higher prevalence of feline DM in recent years is most likely caused by the rise in obesity in our cat population. Obesity considerably increases the risk to become diabetic about 4 times compared to lean cats, and at least 60% of obese cats seem to become diabetic over time (Hoenig, 2006a, 2006b). Further, and similar to humans, the degree of overweight seems to be directly linked to the increased risk of developing DM. In studies by Scarlett and coworkers (Scarlett et al., 1994; Scarlett & Donoghue, 1998), overweight cats were 2.2 times as likely, and obese cats were 6 times as likely to be diabetic than optimal weight cats. Different scoring systems have been described but the most common scoring systems used are the 5-point system (Figure 10) (where a BCS of 3 is considered ideal) or the 9-point system (where a BCS of 5 is considered ideal); (see Obesity chapter).Therefore, any increase in body weight above normal should be avoided to reduce the risk of cats to develop DM (Scarlett & Donoghue, 1998).

Once obesity is established, the heat production and hence the energy requirement, is reduced in obese cats when corrected for metabolic BW (Hoenig et al., 2006c; 2007a, 2007b). This will help to perpetuate obesity unless food intake is rigorously adjusted. In another study (Nguyen et al., 2004a, 2004b), it was reported that total energy expenditure is unchanged in neutered or intact cats of different BW if values are corrected for metabolic BW or for lean body mass. However, Nguyen et al (2004a, 2004b) used a different technique to determine total energy expenditure than Hoenig et al (2007b) which may explain the different outcome.

Obesity and Insulin Resistance

A number of studies have shown that obese cats face a high risk of developing DM because they have a higher baseline insulin concentration, show an abnormal insulin secretion pattern in IVGTT and euglycemic hyperinsulinemic clamp studies, and are insulin resistant (Biourge et al., 1997; Scarlett & Donoghue, 1998; Appleton et al., 2001b; Hoenig et al., 2002; 2007b). Depending on the experimental technique and the degree of obesity, insulin sensitivity was reported to be reduced by 50 to over 80%. Figure 11 shows one example of how glucose tolerance in cats is affected by body weight (see also Figure 13). A cat was considered having abnormal glucose tolerance when glucose half-life was above 80 min in an IVGTT (Lutz & Rand, 1995).

Figure 11. Association between glucose tolerance (assessed by glucose half-life in an IVGTT) and body weight in clinically healthy cats.

Insulin resistance seems to be associated with a decreased expression in the insulin-sensitive glucose transporter GLUT4, while the expression of GLUT1, which mediates insulin-independent glucose transport, is unaltered (Brennan et al., 2004). This effect occurs early in the development of obesity, before overt glucose intolerance is observed. Interestingly, at basal insulin levels glucose utilization seems to be normal in obese cats. However, in a stimulated state (e.g., by IVGTT), not only insulin sensitivity but also glucose effectiveness, that is, the ability of glucose to promote its own utilization at baseline insulin levels, was reduced by approximately 50% (Appleton et al., 2001b; Hoenig et al., 2006c; 2007a, 2007b).

Obesity and Lipid Metabolism

Obese cats have higher baseline concentrations of non-esterified fatty acids (NEFA) than lean cats. This may reflect in part a general change from glucose to fat metabolism in skeletal muscle of obese cats. Lower activity of lipoprotein lipase in body fat combined with higher activity of lipoprotein lipase and of hormone-sensitive lipase in the muscle in obese cats may favor the redistribution of fatty acids from adipose tissue to skeletal muscle (Hoenig et al., 2006b; 2007b). The lipid accumulation in skeletal muscle seen in obese cats could then result in a lower insulin sensitivity because changes in lipid metabolism lead to altered insulin signaling and affect GLUT4 expression (Wilkins et al., 2004; Brennan et al., 2004). In obese cats, both intramyocellular and extramyocellular lipids increase. Whether and how elevated intramyocellular lipids affect GLUT4 expression, and hence insulin sensitivity directly remains to be study. All in all, general obesity clearly favors the development of insulin resistance in muscle (Wilkins et al., 2004).

The link between obesity and the changes in metabolic handling of nutrients in adipose and skeletal muscle tissue may be represented by differential expression of tumor necrosis factor-alpha (TNFα). TNFα reduces lipoprotein lipase, and a study has shown that TNFα is upregulated in adipocytes, but downregulated in skeletal muscle of obese cats (Hoenig et al., 2006b).

TNFα is one of the numerous hormones and cytokines that are released by adipose tissue and that are now considered of pivotal importance for regulating nutrient handling (for review, see Lazar, 2005). All endocrine factors released from adipose tissue are collectively called adipokines. TNFα in particular is not only produced by adipocytes, but also by macrophages. In fact, obesity is considered a low grade inflammatory disease of adipose tissue. Many cytokines released from adipose tissue induce peripheral insulin resistance. For example, TNFα, which is among the best investigated, interferes with insulin signalling and causes insulin resistance.

Adiponectin is the only adipokine known which is inversely related to the amount of body adiposity (for review, see Ahima, 2005). Adiponectin improves insulin sensitivity by increasing fatty acid oxidation, reducing hepatic gluconeogenesis, and by inhibiting inflammatory responses. Because its concentration is reduced in obesity, it combines with increased release of TNFα to promote insulin resistance. However, it has to be pointed out that none of these effects have been investigated in detail in cats (see also Figure 12). It was also claimed that elevated levels of insulin-like growth factor-1 (IGF-1) may constitute the link between obesity and insulin resistance (Leray et al., 2006). However, this has never been shown in cats and the data in other species are also conflicting. Reusch et al (2006a) have shown that diabetic cats have lower IGF-1 levels which increase in response to insulin treatment.

Figure 12. Insulin resistance.

Despite many similarities between human 2DM and feline DM, it should be highlighted that there may also be some distinct differences. One of them being that in cats, insulin suppresses the serum concentration of NEFA’s more in obese than in lean cats. This appears to be due to an increased sensitivity to insulin-induced fatty acid uptake (Hoenig et al., 2003). Further, obese cats seem to accumulate similar amounts of subcutaneous and visceral fat. This may be of importance because in humans, visceral fat in particular has been associated with the metabolic derangements of obesity.

Reversibility of Insulin Resistance

Regarding the possible treatment outcome for diabetic cats, it is important to note that insulin resistance induced by obesity in cats is reversible after the correction of body weight () (Biourge et al., 1997). Hence, if diabetic cats are obese, lowering their body weight to normal should always be part of the therapy. In the course of the above mentioned study (Biourge et al., 1997), cats were also exposed to a poorly palatable diet which resulted in a voluntary decrease in food intake. The ensuing rapid body weight loss led to a deterioration of glucose tolerance and severely depressed insulin secretion. This was, however, temporary. Presumably, insulin resistance was caused by an adaptation to nutrient deprivation and a shift from carbohydrate to fat catabolism. This may result in elevated levels of triglycerides and free fatty acids. Hence, these are increased in obesity, but also during massive caloric restriction and must be considered a normal metabolic adaptation (see also Banks et al., 2006).

Figure 13. The effect of body weight gain and recovery to normal body weight on plasma insulin levels (Biourge et al., 1997).

Even though the phenomenon of increased body weight in neutered cats has been known for a long time, more in-depth studies on underlying causes have only recently been performed. The increase in body weight, and hence the decrease in insulin sensitivity, in cats after neutering appears to result from both an increase in food intake and a decrease in energy requirement (Root et al., 1996; Biourge et al., 1997; Fettman et al., 1997; Harper et al., 2001; Hoenig & Ferguson, 2002; Kanchuk et al., 2002; Kanchuk et al., 2003). The latter effect, however, has been disputed because it was not consistently observed in male cats (Kanchuk et al., 2003). The different outcome of studies may be due to procedural differences. Kanchuk et al (2003), determined energy expenditure as expressed per lean body mass. This was done on the understanding that BW gain in overfed cats results mainly from an increase in adipose tissue mass which is metabolically relative inactive (Kanchuk et al., 2003; see also Martin et al., 2001). In any case, neutered cats have a much higher risk of becoming obese.

General Concepts of Glucotoxicity, Lipotoxicity, and Glucolipotoxicity

Glucose sensing in the feline pancreas seems to be similar to other species. Via the pathways outlined in Figure 6 & Figure 7, glucose and free fatty acids (or NEFA) normally increase insulin secretion. Glucose also promotes normal expansion of beta-cell mass, and the two mechanisms, glucose stimulation and uptake via GLUT2, and glucose-induced cell proliferation seem to be directly linked through distinct intracellular signaling pathways (reviewed in Prentki & Nolan, 2006). The effect of glucose on beta-cell proliferation is further stimulated by incretins such as GLP-1 and free fatty acids. Hence, GLP-1 protects beta-cells from apoptosis and promotes beta-cell growth.

As reviewed by Prentki et al (2002), glucose concentrations below 10 mmol/L (180 mg/dL) normally are not toxic to the pancreatic beta-cells. This refers to physiological postprandial hyperglycemia which triggers beta-cell proliferation (Donath et al., 2005). Similarly, physiologically elevated fatty acid concentrations alone are not toxic, at least when malonyl-CoA, which is a side product of glucose metabolism in beta-cells and which inhibits uptake of fatty acids in mitochondria for subsequent beta-oxidation, is low. Fatty acids increase insulin secretion via increases in Ca2+ and diacylglycerol (Figure 7). Problems only arise when hyperglycemia and elevated fatty acids occur simultaneously and for prolonged periods. While insulin secretion initially is increased via glucose and long chain fatty acid-CoA (Figure 6 & Figure 7), a marked elevation of glucose, and activated fatty acids and further lipid signalling molecules reduce insulin secretion and promote apoptosis. These effects are called glucotoxicity and lipotoxicity, respectively. Because lipotoxicity is most apparent under prevailing hyperglycemia, the term glucolipotoxicity has been coined (Prentki & Nolan, 2006).

Glucotoxicity and Lipotoxicity

The concept of glucotoxicity, or better glucolipotoxicity, is not novel (Rossetti et al., 1990) but research over the last few years has yielded good progress in the understanding of underlying causes and mechanisms. Glucotoxicity and lipotoxicity refer to a defect in stimulus- secretion coupling which ultimately leads to beta-cell failure. Both phenomena occur relatively rapidly so that hyperglycemia sustained for only a few days downregulates the glucose transport system, and an elevation of free fatty acids for 24 hours reduces insulin secretion.

It has to be made clear that only few aspects of gluco- and lipotoxicity have been studied in cats so far. Nonetheless, the author believes that due to the many similarities between rodent models of 2DM and especially human 2DM and feline DM (Henson && O’Brien, 2006), many aspects discussed in the following section are probably also valid for cats (see below).

The reduction in beta-cell mass caused by chronic hyperglycemia and glucotoxicity results from an imbalance between beta-cell neogenesis and proliferation, and beta-cell apoptosis (Donath et al., 2005). During chronic hyperglycemia and hyperlipidemia, glucose, saturated fatty acids and triglycerides accumulate in beta-cells, triggering the release of cytokines. All these factors reduce insulin secretion and favor beta-cell apoptosis. At the cellular level, glucotoxicity is associated with mitochondrial dysfunction which, due to enhanced oxidative glucose metabolism, may be linked to increased oxidative stress in pancreatic beta-cells (Prentki & Nolan, 2006). Reactive oxygen species can be "detoxified", but this happens at the expense of ATP and hence lower insulin secretion (Figure 6 & Figure 7).

Interestingly, the first report on glucotoxicity in cats by was published in 1948. (© Y. Lanceau/RC).

Dysfunctional lipid metabolism, triglyceride and free fatty acid cycling also contribute to beta-cell failure. This results in the accumulation of long chain fatty acid-CoA which directly influences the ATP-sensitive K channel that is involved in glucose-stimulated insulin release. Further, elevated intracellular malonyl-CoA levels reduce the uptake of fatty acids into mitochondria and thereby shift fat metabolism from fatty acid oxidation to fatty acid esterification and lipid accumulation. This results in a lower production of intracellular ATP which is important for stimulus-secretion coupling (Prentki & Nolan, 2006).

In recent years, evidence has also accumulated that glucotoxic and lipotoxic events are directly linked to islet inflammation. Among other factors, interleukin 1-beta (IL-1beta) has been identified as one of the key molecules (Donath et al., 2005). Even though IL-1beta upregulation has now been reported in several animal models of 2DM, further studies are clearly required to investigate the link between hyperglycemia and inflammation (Prentki & Nolan, 2006). The author is not aware of any such studies having been performed in cats to date.

Gluco- and Lipotoxicity in Cats

In their paper entitled Experimental diabetes produced by the administration of glucose, Dohan and Lukens (1948) described the effect of sustained hyperglycemia on the islets of Langerhans. They report that cats developed degranulation of beta-cells followed by degeneration of islets. Many cats developed overt diabetes mellitus, at that time characterized by massive glucosuria.

Glucotoxicity

Glucotoxicity clearly contributes to beta-cell failure in cats but it is reversible if hyperglycemia resolves. However if maintained, permanent loss of beta-cells may ensue. In healthy cats, sustained hyperglycemia of about 30 mmol/L (540 mg/dL) induced by chronic glucose infusion almost completely shut down insulin secretion three to seven days after the start of infusion. Pancreatic histology revealed massive changes in beta-cell morphology. Pancreatic beta-cells showed vacuolation, glycogen deposition, loss of insulin staining and pyknosis. However, even profound histological changes appeared to be reversible upon early resolution of hyperglycemia (Rand & Marshall, 2005). The author’s unpublished studies also clearly show that hyperglycemia of about 25 mmol/L (450 mg/dL) for only 10 days is sufficient to cause a massive decrease in the insulin secretory capacity of pancreatic beta-cells in healthy cats.

Lipotoxicity

Lipotoxicity has not been investigated in detail in cats. However, Hoenig (2002) hypothesized that lipotoxicity might also play a pathogenic role in the diabetic cat. As first described in the glucose fatty acid cycle (Randle cycle; Randle, 1998), glucose inhibits fatty acid oxidation, and vice versa (Figure 14). Because NEFA concentrations are elevated in obese cats and because obese cats are most prone to developing diabetes mellitus, it is plausible to suggest that NEFA reduces glucose metabolism in betacells. However glucose metabolism is a necessary component in glucose-stimulated insulin release. Hence, glucose- stimulated insulin release would be decreased. A study by the same group has shown that saturated fatty acids in particular seem to be detrimental to glucose control in cats while polyunsaturated fatty acids (3-PUFA) may have beneficial effects (Wilkins et al., 2004).

Figure 14. Simplified concept of the glucose fatty acid cycle (Randle cycle; Randle, 1998).

Similar cellular mechanisms as just described for the pancreatic beta-cell also seem to play a role in glucolipotoxicity in insulin target tissues. This has been investigated in less detail but as mentioned earlier, intramyocellular lipid accumulation in skeletal muscle cells reduces their insulin sensitivity (Wilkins et al., 2004; see also Hoenig, 2002). Hence, elevated glucose levels and perturbed lipid metabolism in diabetic cats not only lead to beta-cell failure but may also reduce insulin sensitivity in insulin-target tissues.

All in all, gluco- and lipotoxicity seem to be phenomena which contribute to the progressive deterioration of metabolic control in diabetic cats, both via an effect on pancreatic beta-cells and via an effect on insulin-sensitive target tissue. This clearly underlines the pivotal importance of glucose lowering strategies to curtail this progressive deterioration. Hence, early reversal of hyperglycemia, preferentially by aggressive insulin treatment, reverses glucolipotoxicity, and this will help to achieve diabetic remission in a large number of diabetic cats (see also paragraph on transient diabetes; Nelson et al., 1999).

Amylin as a Circulating Hormone in the Development of Feline Diabetes Mellitus

As discussed, amylin is a normal secretory product of pancreatic beta-cells in all species. Amylin is co-synthesized and co-secreted in parallel with insulin in response to appropriate stimuli (Lutz & Rand, 1996). Hence, changes in plasma insulin levels are usually associated with corresponding changes in plasma amylin levels. In human 2DM and in feline DM, the hormonal situation changes over the course of the disease. Early phases of feline 2DM or mild forms of the disease are often characterized by (compensatory) hyperinsulinemia and absolute or relative hyperamylinemia (O’Brien et al., 1991; Lutz & Rand, 1996). Early hyperamylinemia may favor the deposition of feline amylin as pancreatic amyloid (see below). Progressive beta-cell failure in more severe forms and late stages of feline DM, however, leads to overt hypoinsulinemia and hypoamylinemia (Johnson et al., 1989; Ludvik et al., 1991). Most clinical cases of feline DM are probably presented to veterinarians at that stage.

The regulation of nutrient metabolism by amylin involves modulation of pancreatic glucagon release, the regulation of gastric emptying (for review: Edelman & Weyer, 2002), and an inhibition of food intake (Lutz, 2005). Hence, the lack of amylin in DM results in oversecretion of glucagon, accelerated gastric emptying and overeating. At least in humans and rodents, amylin has been shown to decrease excessive postprandial hyperglucagonemia observed in DM (Fineman et al., 2002) and to normalize gastric emptying. Hyperglucagonemia is also present in diabetic cats (Figure 15; Tschuor et al., 2006), but it is unknown at present whether this is due to the lack of amylin in these animals. However, preliminary studies in healthy cats show a trend for an effect of amylin to reduce glucagon output (Figure 16; Furrer et al., 2005). Similar studies in diabetic cats have not been performed yet. Further, it has not been investigated in detail whether, similar to humans or rodents, gastric emptying in diabetic cats is accelerated. Hence, it is unknown if presuming that such defect were present, this would be due to amylin deficiency.

Figure 15. Baseline hyperglucagonemia in diabetic cats after 12h of fasting (Tschuor et al., 2006).

Figure 16. Amylin slightly reduces measured glucagon blood levels in an arginine stimulation test (AST; Figure 16a) and a meal response test (MRT; Figure 16b) (Furrer et al., 2005).

In summary, there is reason to believe that the lack of amylin in diabetic cats contributes to metabolic dysregulation. The most prominent effect in this regard is the lack of amylin’s suppression of prandial glucagon secretion. Amylin replacement is now a common form of therapy in human DM but is so far unknown in the treatment of diabetic cats.

Pancreatic Glucagon as a Circulating Hormone in the Development of Feline Diabetes Mellitus

Pancreatic glucagon as a pathogenic factor in the development of DM has been neglected for many years due to the overwhelming importance that was given to insulin deficiency as the critical factor. Notwithstanding, deficient suppression of glucagon secretion, especially in the immediate postprandial period, seems to be a major contributor to postprandial hyperglycemia (Figure 16) (O’Brien et al., 1985; Furrer et al., 2005; Tschuor et al., 2006). Diabetic hyperglucagonemia seems to be directly linked to amylin deficiency and hence disinhibition of glucagon release. This may also be true for the cat (Figure 16) (Furrer et al., 2005). To what extent reduced insulin suppression of glucagon release also contributes to the phenomenon in cats, remains to be determined.

Pancreatic Amyloidosis

The most common and consistent morphological feature is islet amyloidosis (Figure 17a & Figure 17b) (Yano et al., 1981; O’Brien et al., 1985; Johnson et al., 1986; Johnson et al., 1989; Lutz et al., 1994; Lutz & Rand, 1997). Amyloid deposition is found in a large proportion of overtly diabetic cats and cats with impaired glucose tolerance, a state also referred to as pre-diabetic (Johnson et al., 1986; Westermark et al., 1987; Lutz & Rand, 1995). Islet amyloidosis is thought to play an important role in the pathogenesis of 2DM and feline DM because it contributes to progressive beta-cell loss which is typically observed over the course of the disease (Höppener et al., 2002).

Figure 17a. Pancreatic islet of a cat with massive deposition of islet amyloid which consists mainly of precipitates of the beta-cell hormone amylin. (© Thomas Lutz).

Figure 17b. The pancreatic islet of a healthy control cat is shown for comparison. Immunohistochemical stain for amylin. Intact beta-cells stain in red, islet amyloid stains in pink. (© Thomas Lutz).

Pancreatic amyloid deposits consist mainly of amylin, hence amylin’s other name islet amyloid polypeptide, or IAPP (Westermark et al., 1987). Pancreatic amylin has the propensity to precipitate as amyloid deposits only in a small number of species such as humans, non-human primates and cats (Johnson et al., 1989; Westermark et al., 1987), and only these species naturally develop a 2DM like syndrome. A necessary precondition is a certain amino acid sequence in the middle part of the amylin molecule in humans and cats (but not rats) that is unrelated to amylin’s hormonal action, but predisposes amylin to form insoluble fibrillar aggregates. A second prerequisite appears to be hypersecretion of amylin leading to high local amylin concentrations in pancreatic islets (Cooper, 1994). Especially during early islet amyloid formation, soluble amylin fibril oligomers contribute to beta-cell toxicity and subsequent beta-cell loss (Höppener et al., 2002; Butler et al., 2003; Konarkowska et al., 2006; Matveyenko & Butler, 2006). A third and only poorly defined factor in the development of islet amyloidosis seems to be some malfunction of pancreatic beta-cells leading to aberrant processing of amylin (Ma et al., 1998).

As mentioned, early phases of feline DM are characterized by hyperamylinemia (O’Brien et al., 1991; Lutz & Rand, 1996). This may favor the deposition of feline amylin as pancreatic amyloid. Progressive beta-cell failure in late stages of feline DM leads to low circulating amylin levels (Johnson et al., 1989; Ludvik et al., 1991; Cooper 1994).

Quantitative Aspects of Islet Amyloid in Cats

Being the most prominent histological finding in diabetic cats, it was very interesting to note that islet amyloid deposition also occurs in non-diabetic, healthy cats. Some of these cats appeared to develop relatively large amounts of islet amyloid without obvious clinical signs (Figure 18) (Lutz et al., 1994). The prevalence of pancreatic amyloid increased with age (Figure 19), hence a finding similar to the general observation of an increased prevalence of feline diabetes in older animals. Most important, however, diabetic cats had markedly larger deposits of pancreatic amyloid than healthy cats, and the extent of amyloid deposition seemed to be directly related to the severity of clinical signs in feline DM (O’Brien et al., 1985; Johnson et al., 1989; Lutz et al., 1994). This is also reflected in the association between the amount of pancreatic islet amyloid and the occurrence of glucose intolerance as assessed via glucose half-life in plasma in an IVGTT (Figure 20).

Figure 18. Frequency of islet amyloid deposition in 84 clinically healthy cats.

Figure 19. Islet amyloid deposition increases with age (Lutz et al., 1994).

Figure 20.The amount of pancreatic islet amyloid is positively correlated to glucose T1/2 as determined in an IVGTT (Lutz et al., 1994).

Unfortunately, even though pancreatic islet amyloid is an important factor in the pathophysiology of feline DM, it cannot be assessed under in vivo conditions. Therefore, it is currently not a helpful prognostic marker for the development of the disease.

Studies in transgenic rodents have clearly pointed to an important role of amylin-derived amyloid in the development and progression of 2DM. Small molecular weight, soluble amylin oligomers in species with an amyloidogenic amino acid sequence, are causative for beta-cell apoptosis (for review: see Muff et al., 2004). Nonetheless, the primary events leading to the formation of these cytotoxic oligomers in 2DM remain to be resolved.

The Link Between Hyperglycemia and the Formation of Islet Amyloid

Now that the major pathogenetic factors (gluco-lipotoxicity and amylin-derived islet amyloid) contributing to progressive betacell failure in diabetic cats have been reviewed, it should be noted that it is as yet completely unknown whether and how there may be a link between these factors. However, it seems possible that changes in the intracellular milieu induced by elevated glucose or fatty acid levels (intracellular stress) may create conditions that promote the formation and precipitation of islet amyloid fibrils. The most toxic form to beta-cells are small molecular oligomers of amylin fibrils which are most likely formed early in the disease process. Hence, any therapy aimed at reducing blood glucose levels, and subsequently at reducing the secretory stress on pancreatic beta-cells, as early as possible in the disease process may favor diabetic remission as seen in transient DM (see below).

Reduced Insulin Sensitivity in Diseased Cats

Similar to humans, glucose homeostasis seems to be frequently impaired in cats suffering from various diseases including severe inflammation, malignant neoplasia, sepsis, viral infection, end-stage renal disease, and chronic heart failure. As an underlying cause, a combination of augmented synthesis of pro-inflammatory cytokines and the presence of insulin counter-regulatory hormones has been hypothesized. This has been substantiated in cats with congestive heart failure which have elevated levels of TNFα (Meurs et al., 2002).

Further, stomatitis, pulmonary lesions (Mexas et al., 2006), and urinary tract infections (Jin & Lin, 2005) seem to be more frequent in diabetic cats. Seriously ill cats may show profound stressinduced hyperglycemia. They do not always suffer from concomitant hyperinsulinemia which would be indicative of insulin resistance (Chan et al., 2006).

The exact mechanisms linking disturbed glucose homeostasis and various illnesses in cats are still largely unknown. Various cytokines are most likely involved. A recent preliminary study has shown that a 10-day infusion of lipopolysaccharide, which is a cell wall component of Gram negative bacteria and which causes the release of various cytokines, leads to impaired glucose tolerance (unpublished). It could also be speculated that these disorders are associated with reduced levels of the adipocyte hormone adiponectin which appears to be an important factor in regulating insulin sensitivity in insulin target tissues (Hoenig et al., 2007a). Apart from effects of cytokines on insulin-sensitive tissues, various cytokines directly reduce pancreatic endocrine secretion.

Finally, it should also be recognized that one is faced with a typical chicken and the egg conundrum. On one hand, hyperglycemia in DM reduces the body defense against infection, for example, in the urogenital tract (e.g., Lederer et al., 2003; Bailiff et al., 2006). On the other hand, infection and inflammatory disorders, perhaps through TNFa, are associated with insulin resistance which may ultimately lead to DM (Figure 21).

Figure 21. Self-perpetuation of diabetes mellitus.