To explore the effect of diabetes and insulin resistance on lipoproteins, we will start at the beginning. In this chapter the beginning must be the chylomicron and its synthesis, since without food there would be little interest in diabetes or insulin resistance and in the rural areas of the world where starvation occurs, there is little talk of type 2 diabetes or insulin resistance. The apo B48-containing chylomicron transports both cholesterol and triglyceride from the intestine to the circulation and has a dominant role in distributing fatty acids/triglyceride to the tissues prior to being taken up by the liver [35]. The second function of the chylomicron is to transport cholesterol to the liver, although on the way the cholesterol may be taken up by tissues including the atheromatous plaque where the macrophage sits in waiting with a specific apo B48 receptor [36] as well as VLDL and scavenger receptors [35]. Apo B48 has been demonstrated in plaque by a number of workers [37–40]. Apo B48 is the solubilising protein necessary for the transport of cholesterol and lipid in aqueous solution in humans. The amount of triglyceride available for the chylomicron particle is limitless, the normal gut managing to limit the amount of triglyceride/fatty acids in the stool to under 5 g/day. On the other hand serum cholesterol is very tightly regulated and varies very little throughout one’s lifetime due to a hugely efficient regulatory process. Absorbed cholesterol varies considerably from person to person, and high absorbers of cholesterol have been shown to have low synthesis rates and to be less sensitive to cholesterol lowering with statin therapy [41]. The mechanism that regulates cholesterol absorption in the intestine is complex and both diabetes and insulin resistance have been shown to affect the regulation [42, 43], causing the initiation of the dyslipidemia of insulin resistance and diabetes (Fig. 5.4).

The first step in cholesterol absorption in the intestine appears to be through the multi-transmembrane protein Niemann-pick C1-like 1 (NCP1-L1) which is highly expressed in the jejunum. In humans it is localised to the brush borders of the enterocytes and acts as a unidirectional transporter of cholesterol and non-cholesterol sterols [44, 45]. The mechanism of action of NCP1-L1 has been elucidated. It has been shown that cholesterol promotes the formation and endocytosis of NCP1-L1-flotillin-cholesterol membrane microdomains which is an early step in cholesterol uptake. Zang et al. [46] discovered that it is the N-terminal domain of NCP1-L1 that binds cholesterol. It is interesting that this domain does not bind to plant sterols; thus, it now seems that plasma membrane-bound NCP1-L1 binds exogenous cholesterol, and this binding facilitates the formation of the NCP1-L1-flotillin-cholesterol microdomains that are then internalised into cells through the clathrin adaptor protein 2 pathway. Twenty rare NCP1L1 alleles have been found in the low cholesterol absorbers and appear to impair NCP1-L1 cholesterol uptake through various mechanism ([47, 48]; for review see Calandera [49]). It has been shown that the effectiveness of ezetimibe, which blocks NCP1-L1 and inhibits cholesterol absorption, depends on the NCP1-L1 genotype [50].

There are other transporters of cholesterol; for example, SR-B1 is located both in the apical and basolateral membranes of the enterocyte [51]. Scavenger receptors (SR) are cell surface proteins that can bind and internalise modified lipoproteins. SR-B1, which is involved in cholesterol uptake in the intestine and may play an important part in intestinal chylomicron production, and the fatty acid transporter CD36, which is also involved in the uptake of oxidised LDL, are members of the class B scavenger receptor family [52]. Hiashi et al. [53] investigated gene expression of key proteins involved in the active absorption of dietary fat and cholesterol in response to the development of insulin resistance. They used two models of diet-induced insulin resistance, the fructose-fed hamster and the high-fat-fed mouse. Expression of SR-B1 was increased in both the fructose-fed hamster and the high-fat-fed mouse models of insulin resistance. In CaCo2 adenocarcinoma cell line, SR-B1 over-expression increased apo B100 and apo B48 secretion. The authors conclude that apical or basolateral SR-B1 may have an important role in cholesterol absorption and may play a part in cholesterol over-absorption in insulin-resistant states. SR-B1 in the intestine may play an important role in chylomicron production. Cdc42, a member of the Rho family of small guanidine triphosphatases with numerous functions, has been shown by Xia et al. [54] to interact with NCP1-L1 and to control its movement from the endocytic recycling compartment to plasma membranes in a cholesterol-dependent manner. Glucose-stimulated Cdc42 signalling appears to be essential for second stage insulin secretion [55]. It is probable that in insulin resistance the signalling of NCP1-L1 is disturbed through this pathway, but we have been unable to find any studies in the intestine that have explored the pathway in diabetes/insulin resistance. In animal studies we have demonstrated an increase in cholesterol absorption in diabetes [56]. We then asked the question as to whether diabetes might be associated with an increase in cholesterol absorption through stimulation of NCP1-L1. We demonstrated in animal models of diabetes that NCP1-L1 was upregulated [57], and in diabetic patients we demonstrated an increase in NCP1-L1 mRNA [58], suggesting a mechanism for an increase in cholesterol absorption. In the Psammomys obesus, a model of type 2 diabetes, the animals exhibiting weight gain, hyperinsulinaemia and hypercholesterolaemia, NCP1-L1 protein and gene expression were both significantly reduced in the intestine, and the authors found a lower capacity to absorb cholesterol compared to controls [59]. This may suggest interspecies variation, but it is a surprising finding considering that this animal model of diabetes has been shown to have increased production of intestinal apo B48-containing lipoproteins [60]. Ezetimibe has been shown to bind to the brush border and to NCP1-L1 expressing cells [61]. There is a sterol regulatory element in the promoter and a sterol-sensing domain of NCP1-L1 which appears to regulate cholesterol absorption in response to cholesterol intake. Huff et al. [62] have shown that NCP1-L1 is suppressed in mice given a cholesterol-rich diet and increased in the cholesterol-depleted porcine intestine. The nuclear receptor, peroxisome proliferator-activated receptor (PPAR) δ/β, appears to control the expression of NCP1-L1. Activation by a synthetic agonist of PPARδ has been shown to reduce cholesterol absorption and reduce expression of NCP1-L1 without altering expression of the adenosine triphosphate (ATP)-binding membrane cassette transport proteins ABC G5/G8 [63]. ABC G5/G8 is a transmembrane heterodimer that transports plant sterols and excess cholesterol out of jejunal enterocytes (discussed in greater detail below). Fenofibrate, a PPARα agonist, has been shown to inhibit cholesterol absorption; the mechanism has been shown to be through reduced NCP1L1 transcription by binding to a PPARα response element upstream of the human NCP1-L1 gene. In a human construct, Iwayanagi et al. [64] showed that PPARα positively regulated human NCP1-L1 transcription, and Valasek et al. [65] showed that fenofibrate reduced intestinal cholesterol absorption by PPARα modulation of NCP1-L1. Tremblay et al. [66] have shown that atorvastatin increases NCP1-L1 in the intestine and decreased ABC G5/G8 which leads to an increase in cholesterol absorption. These findings were accompanied by an increase in the transcription factors, sterol regulatory element-binding protein (SREBP) 2 and hepatic nuclear factor (HNF)-4.

Once cholesterol has been transported across the brush border membrane, it faces another regulatory process and may be excreted back into the intestinal lumen rather than being further processed for absorption into the perimesenteric lymphatic circulation. ABC G5/G8 expression is mostly confined to the human small intestine and liver [67]. These two proteins act in tandem to re-excrete both cholesterol and, in particular, non-cholesterol sterols such as plant sterols from the body. Much of the understanding of ABC G5/G8 comes from the rare mutations that cause a defect in ABC G5/G8 and result in high levels of sitosterol in the blood. Beta-sitosterolemia is a condition which manifests itself in children as tendon xanthomas or in young adults as severe CHD with massive accumulation of sterols and stanols in monocyte-derived macrophages [68]. Ma et al. [69] found in an animal model that dietary calcium had a beneficial effect on lipoprotein profile by upregulating the mRNA levels of intestinal ABC G5/G8 and cholesterol-7α-hydroxylase (CYP7A1), whereas it downregulated the intestinal NCP1-L1 and microsomal triacylglyceride transport protein (MTP) due to enhanced biliary cholesterol excretion. Mendes Gonzales et al. [70] investigated the effect of ABC G5/G8 deficiency on lipoproteins in mice. They found that postprandial triglycerides were fivefold higher in the ABC G5/G8^{− −} mice due to a lower fractional catabolic rate with lower post-heparin lipoprotein lipase activities. They also showed that liver triglyceride secretion and intestinal triglyceride secretion were higher and there was a relationship between this and the HOMA index as a measure of insulin resistance. Since diabetes is so frequently associated with dyslipidaemia and atherosclerosis, the ABC translocases became a target for research. Blocks et al. [71] examined mRNA and protein expression of ABC G5/G8 in the intestine of streptozotocin diabetic rats and found significant reduction in expression of both ABC G5/G8. They found that levels were partially normalised on insulin supplementation. We have shown that ABC G5/G8 were reduced by more than 50 % in the intestine of Zucker diabetic fa/fa rats compared with lean rats although these changes did not reach statistical significance [58]. Insulin treatment caused a nonsignificant increase in ABC G5/G8 mRNA. In another study of streptozotocin diabetic rats, ABC G5/G8 were both very significantly reduced in the intestine [57]. There was a negative correlation between ABC G5/G8 and chylomicron cholesterol [57]. In the Psammomys obesus, another model of diabetes, Levy et al. [59, 60] showed a reduction in ABC G5/G8 in the intestine. In the intestine of human subjects with type 2 diabetes, ABC G5/G8 mRNA were both significantly lower compared to controls [58]. There was a negative correlation between ABC G5/G8 and NCP1L1 in the combined diabetic and control subjects, and there was a significant negative correlation between chylomicron cholesterol and both ABC G5/G8 [57]. These two genes appear to play an important role in the dysregulation of cholesterol metabolism in diabetes.

The cholesterol that has evaded ABC G5/G8 in the intestine is now ready to be solubilised for transport through the lymphatic system. The assembly of the chylomicron occurs under the direction of microsomal triglyceride transfer protein (MTP). MTP has the ability to combine cholesterol, triglyceride and phospholipid into the triglyceride-rich chylomicron particle. The cholesterol that becomes available however not only is cholesterol that has been absorbed from the diet but is also cholesterol that has been excreted through the bile duct under the influence of hepatic ABC G5/G8. Finally, there is the cholesterol that has been synthesised in the intestine through the 3-hydroxy-3-methylglutaryl coenzyme A (HMGCoA) reductase pathway. This pathway in the intestine accounts for up to 25 % of body-synthesised cholesterol, the amount varying depending on whether the patients are high or low cholesterol absorbers. Intestinal MTP plays a major role in the assembly of the chylomicron particle and therefore of cholesterol and triglyceride metabolism. MTP has become a hot topic since inhibitors of intestinal MTP have been shown to lower triglyceride without causing hepatic steatosis at least in animal studies [72, 73]. In short-term human studies, a specific intestinal MTP inhibitor did not appear to effect liver function tests [74]. Although many polymorphisms of MTP have been described, some of which have considerable impact on LDL cholesterol in both nondiabetic and diabetic subjects, it is difficult to know whether the results mainly stemmed from the effect in the liver rather than the intestine [75, 76]. The intestinal inhibitors of MTP which have no effect on the liver should answer this question in the future. In animal studies, diabetes is associated with an increase in MTP mRNA with close correlation between MTP mRNA and chylomicron cholesterol [77–80]. In the diabetic rabbit, increased intestinal MTP mRNA is associated with an increase in chylomicron particle numbers [77], but in the rat it is associated with larger particles [78]. The fructose-fed insulin-resistant hamster model had an increase in MTP protein mass, and this was associated with an increase in the triglyceride-rich intestinally derived lipoproteins [79]. Zolotowska et al. [80] in 2003 [80] examined the B48-containing lipoprotein assembly in the small intestine of Psammomys obesus, a model of nutritionally induced diabetes and insulin resistance. De novo triglyceride synthesis, apo B48 biogenesis and triglyceride-rich lipoprotein assembly were all increased. MTP activity and protein expression, however, were not altered. In the enterocyte of fructose-fed golden hamster, MTP mRNA and protein mass were increased by TNFα, but apo B levels in the enterocyte were not effected suggesting that there is considerable interspecies variation [80]. In humans with type 2 diabetes, we demonstrated an increase in MTP mRNA in intestinal biopsies [58, 81]. Diabetic patients who were on statin therapy had lower MTP mRNA compared to those not on statins [81]. We found positive correlations between MTP mRNA and chylomicron fraction cholesterol and apo B48 [81]. A novel intestinal-specific inhibitor of MTP has been shown to ameliorate impaired glucose and lipid metabolism in Zucker diabetic fatty rats, but whether this effect was due to impairment of food intake or to inhibition of fat absorption is not clear [82].

The signals that upregulate chylomicron formation to cope with excess fat in the diet are slowly being elucidated. Another non-specific inhibitor of MTP, which reduced serum levels of triglycerides by more than 70 %, was also associated with significant improvements in glucose tolerance and insulin sensitivity in Zucker fatty rats [83]. Hepatic MTP mRNA expression is negatively regulated by insulin, and it is suggested that insulin might also directly inhibit apo B48 secretion independently of MTP even though it is probable that upregulation of MTP stimulates apo B secretion [84]. The membrane glycoprotein CD36 binds long-chain fatty acids. CD36 deficiency reduces chylomicron production [84]. It has recently been shown that binding of lipid by CD36 upregulates apo B48 and MTP through CD36 signalling via the ERK-1/ERK-2 pathway [85]. Interestingly polymorphisms of MTP which have been associated with differences in serum lipids appear to alter cholesterol absorption but not synthesis in women [86].

Apo B48, the structural protein for the chylomicron, is produced in the intestine by editing of the hepatic version, apo B100 [87]. The enzyme apobec cuts the apo B100 form into the shorter version, apo B48. It has been suggested that apo B is in excess of body needs. In the liver recent work has demonstrated that insulin silences apo B translation by introducing intracellular traffic into mRNA granules [88]. The authors showed that the availability of apo B mRNA for translation was regulated by the rate of release from translationally silenced mRNPs processing bodies (p bodies). Insulin specifically silences apo B mRNA translation by reprogramming its mRNA into p bodies and reducing the size of translationally competent mRNA pools. Translational control via traffic into cytoplasmic RNA granules may be an important mechanism for controlling the rate of apo B synthesis and hepatic lipoprotein production, the authors suggest. It is however not clear that this silencing plays a part in reducing chylomicron production or influences nascent chylomicron size. In diabetes it may be that there is an increase in apo B48 production, but then if meaningful, one would expect smaller chylomicron particles containing less triglyceride per particle to be produced. Our studies in an animal model demonstrated that the particles in the cannulated lymphatic duct of the rabbit were associated with an increase in chylomicron particle numbers [89], but in the rat it was associated with larger particles [56]. In patients with type 2 diabetes, apo B48 is increased, but it is difficult to ascertain whether the increase is due to delayed delipidation, increased synthesis or both [90]. We injected labelled chylomicrons, collected by cannulation of the lymph duct into diabetic and nondiabetic rabbits, into another group of diabetic and nondiabetic rabbits and found evidence of both increased synthesis and delayed clearance [89]. Our animal research therefore suggests that the increase in apo B48 particles in diabetes may be due to both an increase in synthesis and a decrease in turnover.