The relative expression patterns of the two IR (insulin receptor) isoforms, +/− exon 11 (IR-B/IR-A respectively), are tissue-dependent. Therefore we have developed insulin analogues with different binding affinities for the two isoforms to test whether tissue-preferential biological effects can be attained. In rats and mice, IR-B is the most prominent isoform in the liver (>95%) and fat (>90%), whereas in muscles IR-A is the dominant isoform (>95%). As a consequence, the insulin analogue INS-A, which has a higher relative affinity for human IR-A, had a higher relative potency [compared with HI (human insulin)] for glycogen synthesis in rat muscle strips (26%) than for glycogen accumulation in rat hepatocytes (5%) and for lipogenesis in rat adipocytes (4%). In contrast, the INS-B analogue, which has an increased affinity for human IR-B, had higher relative potencies (compared with HI) for inducing glycogen accumulation (75%) and lipogenesis (130%) than for affecting muscle (45%). For the same blood-glucose-lowering effect upon acute intravenous dosing of mice, INS-B gave a significantly higher degree of IR phosphorylation in liver than HI. These in vitro and in vivo results indicate that insulin analogues with IR-isoform-preferential binding affinity are able to elicit tissue-selective biological responses, depending on IR-A/IR-B expression.
- blood glucose
- insulin action
- insulin receptor-A (IR-A)
- insulin receptor B (IR-B)
The incidence of Type 2 diabetes constitutes the majority (>90%) of diabetes cases and is characterized by chronic hyperglycaemia resulting from defects in insulin action, insulin secretion or both. During the last decades insulin analogues, such as fast-acting insulin aspart, lispro and glulisine and long-acting insulin detemir and glargine, have improved glycaemic control and made insulin therapy more convenient for people with diabetes. The development of new insulin therapies has mainly focused on the time–action profile of the glucose-lowering response. However, as the number of people with Type 2 diabetes increases, focus must now be directed not only towards glycaemic control, but also towards other abnormalities associated with the metabolic syndrome. In order to develop improved insulin therapies, a biological approach can be taken.
The IR (insulin receptor) exists in two splice variants, depending on the presence (IR-B) or absence (IR-A) of a small amino acid sequence (encoded by exon 11), which is located at the C-terminal of the extracellular α-subunit . The isoforms are expressed in a highly tissue-specific manner in humans, with IR-B being the dominant isoform in the classical insulin-sensitive tissues, skeletal muscle, adipose tissue and liver, as opposed to IR-A in the brain [2,3]. Additionally, the two isoforms are functionally distinct regarding their binding affinities for insulin, IGF (insulin-like growth factor)-1 and IGF-2 [4,5]. A few groups have tried to address further the biological differences downstream of the two IR isoforms. It has been shown that, in vitro, the isoforms follow different internalization kinetics, with IR-A being internalized faster than IR-B  and, despite IR-B exhibiting a slightly lower affinity for insulin, it might compensate by having increased tyrosine kinase phosphorylation . In addition, Leibiger and co-workers  showed that insulin stimulated its own gene transcription through IR-A, whereas signalling through IR-B resulted in gene transcription of glucokinase in β-cells. Furthermore, IR-A seems to be up-regulated in early developmental stages and in certain cancers [4,9], and it has been shown that IGF-2 especially seems to signal through IR-A and not IR-B [4,9,10], but the exact physiological significance of the two IR isoforms is still not understood.
Tissue-specific IR-knockout studies in mice have shown that ablation of the IR in the liver causes impaired glucose tolerance, hyperinsulinaemia, decreased insulin clearance and an increased risk for the development of atherosclerosis [11,12]. In contrast, specific IR-knockout in adipose tissue [FIRKO (fat-specific IR knockout)] protects against diet-induced obesity and increases lifespan [13,14]. An IR-B-preferential analogue will in theory reach the liver, whereas the expression of the two IR isoforms has to be determined in adipose tissue in different species to understand the prospects of isoform-preferential analogues in adipocytes.
In normal physiology, insulin is secreted into the portal vein and the liver is the first target organ. In the liver, insulin inhibits hepatic glucose production and approximately 50% of the secreted insulin is cleared before reaching the periphery . Thus approximately four times higher levels of insulin reach the liver compared with peripheral target organs, such as fat and muscle. In contrast, exogenous insulin administered by subcutaneous injection results in a non-physiological distribution such that the periphery and the liver are exposed to similar insulin levels . This means that the liver is exposed to lower insulin concentrations and the periphery to relatively higher insulin concentrations than under physiological conditions. Owing to the tissue-specific expression of the IR isoforms, an IR-B-preferential analogue should primarily reach the liver (with an expression level >95%). An insulin analogue mainly targeting the liver is likely to mimic the natural route of insulin distribution, and thereby lead to improved blood glucose regulation by lowering hepatic glucose production. A possible beneficial effect of a liver-specific insulin analogue could be a lower incidence of hypoglycaemia. The glucose transporter in the liver (GLUT2) works independently of insulin levels and is a low-affinity glucose transporter, as opposed to the glucose transporter in the periphery (GLUT4), which is insulin-dependent and a high-affinity glucose transporter . Weight neutrality, by preventing the excess glucose uptake in adipose tissue that otherwise occurs with exogenous administered insulin, could be another beneficial effect of a liver-specific insulin analogue.
The aim of the present study was first to determine the tissue distribution of the IR isoforms and then to examine the biological effects in vitro as well as in vivo of two insulin analogues with different binding affinities for the two IR isoforms.
MATERIALS AND METHODS
Drugs and animals
Insulin analogues and HI (human insulin) were produced at Novo Nordisk by recombinant DNA techniques and site-directed mutagenesis . The insulin analogue INS-A [B(1-29)-VGLSSGQ-A(1-21)-A18Q human insulin] showed increased relative affinity towards the human IR-A isoform, whereas analogue INS-B [A8H,B25N,B27E,desB30 human insulin] showed relatively higher affinity for the human IR-B isoform.
All animal studies were carried out under permits from the Animal Experiments Inspectorate, Ministry of Justice, Denmark. Male Sprague–Dawley and Wistar rats, and male NMRI (Naval Medical Research Institute) and C57bl/6J mice were obtained from Taconic. Unless otherwise stated, animals had free access to food and drinking water. Male LYD (Landrace/Yorkshire/Duroc) pigs were fed a Prima Antonio diet (27501-A; Agrosoft) with free access to water.
Membrane-associated IRs were purified from BHK (baby-hamster kidney) cells, which were stably transfected with a pZem vector containing either the human IR-A or IR-B insert . BHK cells from a ten-layer cell factory were harvested and homogenized in 25 ml of ice-cold buffer (25 mM Hepes, pH 7.4, 2.5 mM CaCl2, 1 mM MgCl2, 250 mg/l bacitracin and 0.1 mM Pefablock). The homogenate was layered on to 41% sucrose cushions, centrifuged (75 min at 95000 g at 4°C) and plasma membranes were collected, diluted 1:4 with buffer and centrifuged again (45 min at 40000 g). The pellets were suspended in assay buffer (50 mM Hepes, pH 7.4, 150 mM NaCl, 5 mM MgSO4 and 0.01% Triton X-100). Membrane proteins were incubated for 150 min at 25°C with 50 pM 125I-TyrA14-labelled HI in a total volume of 150 μl of assay buffer [50 mM Hepes, pH 7.4, 150 mM NaCl, 5 mM MgSO4, 0.01% Triton X-100, 0.1% HSA (human serum albumin) and Complete™ EDTA-free protease inhibitors], 50 μg of WGA (wheat germ agglutinin)-coated PVT (polyvinyltoluene) microspheres and increasing concentrations of HI, INS-A or INS-B (0–300 nM). The assay was terminated by centrifugation (2 min at 750 g) and bound radioactivity was counted on a Packard TopCount NXT γ-counter. Duplicate samples were run for every experiment.
Lipogenesis in isolated primary rat adipocytes
Lipogenesis was determined using isolated primary rat adipocytes as described previously  with minor modifications. In brief, male Sprague–Dawley rats weighing 80–100 g were killed, and the epididymal fat pads were removed and placed in a degradation buffer (110 mM NaCl, 3.3 mM CaCl2, 4.7 mM KCl, 1.2 mM KH2PO4, 1.2 mM Mg2SO4, 11.1 μM dextrose, 25 mM Hepes, pH 7.9, and 4% HSA, adjusted to pH 7.4) containing collagenase at 37°C for 1–1.5 h under vigorous shaking. The cell suspension was filtered, and adipocytes were washed twice and resuspended in an incubation buffer (110 mM NaCl, 4.7 mM KCl, 1.2 mM KH2HPO4, 1.2 mM Mg2SO4, 3.3 mM CaCl2, 25 mM Hepes, pH 7.9, and 1% HSA, adjusted to pH 7.4) and 100 μl aliquots were pipetted into 96-well Picoplates (Packard). Increasing concentrations of insulin or analogue (0–300 nM) were added, and the assay was initiated by the addition of 10 μl of D-[3-3H]glucose and unlabelled glucose to a final concentration of 0.5 mM. The plates were shaken for 2 h and the reaction was stopped by the addition of MicroScint-E (Packard Instrument). The incorporation of radioactively labelled glucose into fat was counted on a TopCounter (Packard Instrument). Triplicate samples were run for every experiment.
Glycogen accumulation in isolated rat hepatocytes
Hepatocytes were isolated from ad libitum-fed male Sprague–Dawley rats (~200 g) by a two-step perfusion technique essentially as described by Seglen . Cell viability, measured using a NucleoCounter (Chemometec), was consistently greater than 90%. Cells were cultured as described previously . For glycogen accumulation experiments, the medium was changed after 24 h to basal medium supplemented with 9.5 mM glucose, 0.1% HSA and different concentrations of insulin or insulin analogues (0–300 nM). The experiments were terminated after 24 h and the glycogen accumulation per mg of protein was determined by amyloglucosidase (exo-1,4-α-D-glucosidase, EC 126.96.36.199) digestion as described by Gómez-Lechón et al. .
Muscle glycogen synthesis
Male Wistar rats (~50 g) were killed by cervical dislocation and the skeletal muscle EDL (extensor digitorum longus) was gently dissected free. Intact EDL muscles were incubated free-floating in pre-gassed (95% O2/5% CO2) KRH (Krebs–Henseleit) buffer (118.5 mM NaCl, 4.7 mM KCl, 1.2 mM KH2PO4, 25 mM NaHCO3, 2.5 mM CaCl2, 1.2 mM MgSO4 and 10 mM Hepes, pH 7.4), supplemented with 0.1% HSA and 5 mM glucose. Following 30 min of pre-incubation, muscles were incubated in KRH buffer supplemented with 0.1% HSA and 5 mM D-[U-14C]glucose with increasing concentrations of insulin or insulin analogues for 30 min. The muscles were then quickly blotted on to filter paper, freeze-clamped in liquid nitrogen, weighed and boiled in 1 M NaOH for 30 min. Glycogen was precipitated overnight at −20°C in ethanol after the addition of 0.35 mg/ml unlabelled glycogen (Sigma). After centrifugation (20 min at 2800 g), the glycogen pellet was washed in ice-cold ethanol, solubilized in water and the radioactivity was measured by liquid-scintillation counting (TRI-CARP 1500; Packard Instruments). All incubations were performed at 30°C, with gentle agitation (110 rev./min), under continuous gassing with 95% O2/5% CO2.
mRNA expression of IR isoforms analysed by multiplex RT (reverse transcription)–PCR
Tissues from three male C57bl/6J mice (~25g), three male Sprague–Dawley rats (~250) and four male LYD pigs (~46 kg) were removed and frozen in liquid nitrogen. The tissues collected from the mice included brain, adipose tissue (epididymal, mesenteric and retroperitoneal), liver, kidney, spleen, heart and two skeletal muscles (EDL and soleus muscles). For the rats, the same tissues were removed as for the mice with the inclusion of subcutaneous adipose tissue and the exception of the spleen. For the pigs, the tissues removed included brain, adipose tissue (mesenteric, retroperitoneal and subcutaneous), liver, kidney, spleen, heart and three skeletal muscles (trapezius, EDL and soleus muscles).
The multiplex RT–PCR method, described previously in detail , allows the co-amplification of several cDNA products from total RNA preparations in a single tube. Briefly, total RNA from mouse, rat and pig tissues as well as primary cells was isolated using RNeasy Mini Kit (Qiagen). The cDNA was synthesized from 1 μg of RNA using an iScript cDNA synthesis kit (Bio-Rad Laboratories). PCR was performed for 28 cycles for all primer sets used. Primers for the IR distinguishing +/− exon 11 included mouse FP (forward primer), 5′-AATCAGAGTGAGTATGACGAC-3′ and RP (reverse primer), 5′-TGTGCTCCTCCTGACTTGT-3′; rat FP, 5′-ACCTTCGAGGATTACCTGCAC-3′, and RP, 5′-GTCTCAGGCCAGAGATGACA-3′; and pig FP, 5′-CTTTTGAGAGGTGGTGAAC-3′ and RP, 5′-ATACAGCACGATCAGACCAT-3′. As an internal control, the housekeeping gene Nono (Non-POU-domain-containing octamer-binding protein) was used because of its stable expression  within the same linear range as the IR. For all species the same primer pair was used (FP, 5′-TGCCAAAGTGGAGCTGGAC-3′ and RP, 5′-ACAATGACTACAGCCCTCTC-3′). Products of the multiplex RT–PCR were resolved by PAGE on a 6% denaturing gel. These were dried an exposed overnight to a Storage Phosphor screen. Band intensity was measured using a STORM840 (Molecular Dynamics) and ImageQuant 5.2 software.
In vivo tissue-differential IR activation in mice
Food was removed from male NMRI mice (25 g) 2 h prior to the insulin injections. Insulin and the insulin analogues were formulated in a neutral buffer solution (5 mM sodium phosphate, 140 mM NaCl and 70 p.p.m. polysorbate-20, pH 7.4) and dosed intravenously in the tail vein (5 ml/kg of body weight). Human insulin and the insulin analogue INS-A were dosed at 6 nmol/kg of body weight. As determined in a previous pilot study in mice, the analogue INS-B lowered blood glucose to a somewhat lesser extent than HI and therefore it was given in a higher dose (9 nmol/kg of body weight). The aim was to obtain similar blood-glucose-lowering in all groups. At various time points, blood samples for blood glucose concentration determination were taken from the tail tip, followed by cervical dislocation and removal of the liver, hindlimb muscle and epididymal fat. Samples were frozen in liquid nitrogen.
For IR phosphorylation analysis, the frozen tissues were homogenized [1:24 (w/v) for muscle, 1:20 (w/v) for liver and 1:3 (w/v) for fat] in ice-cold buffer containing 10 mM Tris/HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM sodium fluoride, 20 mM sodium pyrophosphate, 2 mM sodium vanadate, 1% Triton X-100, 10% glycerol, 0.1% SDS, 0.5% deoxycholate and 1 mM AEBSF, as well as 50 μl of proteinase inhibitor cocktail (P-2714; Sigma) per ml of buffer. Homogenates were cleared by centrifugation (15 min at 12000 g at 4°C). The protein content in the supernatant was measured using the bicinchoninic acid method (Pierce). Protein lysate (60 μg) from liver, skeletal muscle and adipose tissue was loaded on to a phospho-specific ELISA plate recognizing phospho-Tyr1162/1163 at the β-subunit of the IR (KHR9131; Invitrogen), following the manufacturer's instructions.
All results are means±S.E.M. Statistical significance was calculated using an unpaired Student's t test and P<0.05 was considered significant. Data from the different biological in vitro assays were analysed according to the four-parameter logistic model , assuming constant slope, basal and maximal response. The potencies of the two analogues in a given assay are expressed relative to that of HI [EC50(insulin)/EC50(analogue)×100%] measured within the same experiment.
The insulin analogue INS-A has a relatively higher binding affinity for IR-A, whereas INS-B has a higher affinity for IR-B
Years of research within insulin analogue engineering has indentified two insulin analogues, INS-A and INS-B, with different binding affinities for the two human IR isoforms IR-A and IR-B respectively. The competition curves for displacement of 125I-labelled insulin by HI, INS-A and INS-B (Figure 1) clearly show that INS-A has a higher affinity for IR-A than INS-B, whereas INS-B has a higher affinity for IR-B than INS-A. This opens the possibility for applying tissue-selective insulin therapy, since IR-A and IR-B have tissue-specific expression patterns. In order to verify the biological effects in vitro and in vivo of the findings in the binding assay, the tissue distribution of IR-A and IR-B was determined in different animal species.
Tissue distribution of IR-A and IR-B is not conserved among species
The tissue distribution of the IR isoforms in mice, rat and pigs was analysed by multiplex RT–PCR. Mice predominantly expressed IR-A in brain, spleen and two skeletal muscles (EDL and soleus muscles) (Figure 2). IR-B was the dominant isoform expressed in epididymal adipose tissue, as well as in liver and kidney. In the mesenteric adipose tissue, retroperitoneal adipose tissue and heart, a mixture of the two isoforms was found. The same tissue distribution of IR-A and IR-B was found in rats. In addition, a mixture of the two isoforms was found in subcutaneous adipose tissue, which was not included in the analysis of the mice tissues.
In pigs, IR-A was the dominant isoform in brain, spleen and the three adipose tissues (subcutaneous, mesenteric and retroperitoneal), whereas IR-B was the isoform primarily expressed in liver, heart and the three skeletal muscles (trapezius, EDL and soleus muscles). The presence of IR-A in the fat tissue and IR-B in the skeletal muscles of the pig is opposite to the expression in rodents and, therefore, it has to be pointed out that the mRNA expression pattern of the two isoforms are not conserved among species, with the exception of liver and brain, which predominantly express IR-B and IR-A respectively.
INS-A and INS-B have isoform-preferential responses in vitro
Rat liver and epididymal fat predominantly expressed the IR-B isoform. Therefore the effects of the two analogues on glycogen accumulation and lipogenesis were tested in cultured rat hepatocytes and adipocytes. Rat hepatocytes were incubated for 24 h in the presence of HI or insulin analogues and the accumulation of glycogen was measured (Figures 3A and 3B). As shown, it is clear that, compared with HI, INS-B has a higher potency than INS-A for glycogen accumulation in rat hepatocytes. Results obtained from lipogenesis in rat adipocytes showed that INS-B had the same potency as HI, whereas INS-A was less potent than HI (Figures 3D and 3E). Furthermore, as shown in Figures 3(C) and 3(F), hepatocytes as well as adipocytes in culture almost exclusively expressed IR-B, in accordance with in vivo mRNA data. This explains the reduced ability of INS-A to induce glycogen accumulation and lipogenesis, since INS-A has a reduced affinity for IR-B.
In order to examine the biological effect of the two insulin analogues in a system where the IR-A isoform dominated, insulin-stimulated glycogen synthesis in rat muscle strips was measured. The relative expression of IR-A and IR-B at mRNA level in EDL muscles in rats demonstrated that IR-A is the dominant isoform (Figure 4C). As shown in Figures 4(A) and 4(B), both analogues had a lower potency relative to HI. However, as summarized in Table 1, INS-A was found to be more potent in muscles (26%) compared with 5 and 4% in liver and fat respectively, whereas the opposite was observed for INS-B. This suggests that the biological effect of isoform-preferential insulin analogues as expected depends on the dominant isoform expressed in a given cell. However, as also shown in Table 1, INS-B had a higher potency (45%) than INS-A (26%) in stimulating glycogen synthesis in the rat muscle strips, suggesting that perhaps INS-B in general had a higher potency than INS-A in the in vitro systems examined. However, as already mentioned, it is notable that INS-A has a higher potency in muscles than in adipocytes or hepatocytes (opposite to INS-B), suggesting that, at least in vitro, the insulin analogues are isoform-preferential.
Isoform-preferential insulin analogues also show isoform-selective responses in vivo
Both analogues showed isoform-preferential biological responses in vitro and in order to examine whether this was also the case in vivo an acute study of intravenously administered HI or insulin analogues was established. Male NMRI mice were given doses of insulin analogue (6 nmol/kg of body weight for INS-A and 9 nmol/kg of bodyweight for INS-B) to obtain blood-glucose-lowering comparable with that of HI (6 nmol/kg of body weight). At the given doses, both analogues lowered blood glucose similarly compared with HI (Figures 5A and 5E). The three classical insulin-sensitive tissues, liver, fat and skeletal muscles, with distinguished IR-A/IR-B mRNA expression were examined for insulin action at various time points. Insulin action was determined as the analogues ability to phosphorylate the IR at Tyr1162/1163 in liver, hindlimb muscles and epididymal adipose tissue compared with HI.
INS-A appeared less efficient than HI in phosphorylating the IR in the liver, although this was not statistically significant (P=0.07). No significant differences were observed between INS-A and HI in muscles, whereas INS-A was significantly (P<0.05) less efficient in IR phosphorylation in adipose tissue 5 and 15 min after intravenous injections (Figures 5B–5D). Furthermore, INS-B was significantly (P<0.01) more efficient in phosphorylating the IR in the liver compared with HI (Figure 5F). In adipose tissue, there seems to be a tendency (not significant, P=0.052) towards INS-B being slightly better or equipotent with HI in phosphorylating the IR after 5 min (Figure 5H), whereas no differences between HI and INS-B were observed in the muscles (Figure 5G). These findings support the principle that isoform-preferential insulin analogues also have isoform-preferential responses in vivo and that INS-B is liver-preferential when administered acutely.
The present study was carried out to examine the biological difference between two IR-isoform-preferential analogues. Our results show that the IR-A/IR-B-preferential insulin analogues have tissue-specific responses both in vitro and in vivo, reflecting the isoform expression pattern.
It has been argued that the mRNA expression of the two IR isoforms is tissue-dependent and conserved among species in the classical insulin-sensitive tissues, liver, muscle and fat [2,3]. Our present results suggest that this tissue distribution is not completely conserved among species. We confirmed that there is conservation of the predominant expression of IR-B in the liver and IR-A in the brain in mice, rats and pigs, but we found that the relative expression of IR-A and IR-B in skeletal muscle and adipose tissue is not conserved, in agreement with findings in humans , rhesus monkeys  and sheep . Thus perhaps the well-accepted notion that the IR isoform tissue distribution is well-conserved among species and therefore more likely to have functionally distinctive roles may have to be revised and explored further. However, the expression of IR-B in the liver and IR-A in brain and spleen seems to be highly conserved. Whether the mRNA expression corresponds to the protein level was not examined, since several antibodies, claimed to be isoform-specific, were tested and found to be ineffective in detecting proteins by Western blot analysis. However, other studies have shown that the ratio of mRNA expression of the isoforms corresponds to the detected protein level [29,30], despite the fact that the isoforms follows different internalization kinetics  and that they may be more or less likely to form hybrids with the IGF-1 receptor [19,31].
The clear contrast in the IR-A and IR-B mRNA expression pattern in rodents among the primary insulin target organs, with IR-B highly expressed in liver and fat and IR-A highly expressed in skeletal muscle, was exploited in order to test whether IR-isoform-preferential insulin analogues could elicit differing biological responses both in vitro and in vivo. The two examined insulin analogues, INS-A and INS-B, showed isoform-preferential binding, with INS-A having a 5:1 relative higher affinity for human IR-A compared with IR-B (55 compared with 10% relative to HI; see Table 1) and INS-B with the relative ratio 3:1 for human IR-B compared with IR-A (28 compared with 9% relative to HI). Both analogues were normalized to HI, which is known to have a slightly higher affinity for IR-A than IR-B . However, the values given in Table 1 are a ratio (potency) between the EC50 values of each single assay, since this ratio proves to have less variation, reflecting that the effect of the two insulin analogues is the same from assay to assay, despite differences in the absolute EC50 values. The major aim of the present study was to examine whether IR-isoform-selective insulin analogues would have correspondingly biological responses depending on the expression of the isoforms. Since the IR-A and IR-B expression pattern observed in rat hepatocytes, adipocytes and muscle strips corresponded to the expression patterns determined in the intact organs (Figures 2, 3C, 3F and 4C), biological responses were examined in these three cell systems. We found that INS-A exhibited higher potency for muscle glycogen synthesis compared with lipogenesis in adipocytes and glycogen accumulation in hepatocytes, whereas the opposite was observed for INS-B (see Table 1). Surprisingly, INS-B showed higher (45%) potency in stimulating glycogen synthesis in rat skeletal muscles than INS-A (26%), despite IR-A being the predominantly isoform expressed in the muscles. This suggests that, even though there seems to be a clear effect on biological response depending on isoform expression, it looks like that INS-B in general had a higher potency than INS-A in the in vitro systems examined. However, similar to that observed in the cell systems (hepatocytes, adipocytes and muscle strips), intravenous doses that gave a similar glucose-lowering effect showed that INS-B was significantly better than HI in stimulating the phosphorylation of the IR in the liver, whereas INS-A demonstrated a tendency towards stimulating IRs in the liver slightly less than HI, although not significantly. Since the objective of the in vivo experiment was to validate the phosphorylation efficacy of the two insulin analogues at similar blood glucose levels, a higher dose of INS-B than INS-A was required. This is in accordance with an expected lower availability of INS-B to the periphery as relatively more INS-B will be cleared by the liver. Food was only removed 2 h before the experiment and therefore the blood glucose level was mainly determined by the muscle glucose uptake and not by inhibiting hepatic glucose production [32,33]. The in vivo results, together with the differences in potency observed in the three cell systems, indicate that it is possible to obtain a tissue-preferential IR activation, resulting in a tissue-differential biological response using isoform-selective insulin analogues.
Exogenously administrated insulin by subcutaneous injection results in a non-physiological distribution such that the periphery and the liver are exposed to similar insulin levels . The periphery and especially muscle relies on insulin-stimulated GLUT4 translocation for glucose uptake . Since GLUT4 is regulated by insulin independent of blood glucose concentrations  it is likely that a liver-preferential insulin analogue will lead to fewer events of hypoglycaemia. This is of high importance since the optimal insulin treatment for diabetic subjects is often limited because of the risk for hypoglycaemia  and therefore it can be difficult to obtain recommended HbA1c (glycated haemoglobin) targets of <7% . In humans, IR-B is highly expressed in the liver and therefore it may be possible to obtain a more liver-preferential effect with an IR-B-preferential analogue. Since the liver is a glucose-sensitive organ which can switch to glucose production in response to counter-regulatory mechanisms activated by low glucose , it is likely that an IR-B-preferential insulin analogue targeting the liver would be expected to give less hypoglycaemic incidences and thus allow for tighter glucose control. However, it is questionable whether an IR-B-specific analogue will be liver-specific in humans. In humans, in contrast with rodents, IR-B is the predominant isoform in skeletal muscle (~70%) [2,29,37,38] and therefore the IR-B selectivity as a clinical approach to obtain liver specificity may have to be refined further. Another important point to highlight is that previous studies have shown that the isoform distribution changes with age in rats at least  and whether this also occurs in human tissues needs to be examined.
If possible, another beneficial effect of a liver-preferential analogue may be weight-neutrality by avoiding the periphery and especially excess glucose uptake in the adipose tissue. This is supported by the observation that FIRKO mice have slightly increased insulin sensitivity and are resistant to diet-induced obesity , which highlights the potential of an insulin analogue to avoid adipose tissue. The in vitro and in vivo results of INS-A showed that this analogue was less potent in adipose tissue compared with HI and, therefore, an IR-A-preferential analogue may be a different way to obtain weight-neutrality. Only the epididymal adipose tissue was analysed with regard to receptor phosphorylation. In humans, subcutaneous fat expresses 70% IR-B (S. G. Vienberg, K. Højlund and J. F. Wojtaszewski, unpublished work), but so far there are no results on visceral fat, which is known to be the most metabolically active adipose tissue . If visceral fat in humans expresses the IR-A isoform in contrast with IR-B in the subcutaneous fat, this would be an advantage for storing lipids in the more healthy subcutaneous fat and thereby preventing visceral obesity.
In summary, we have shown that insulin analogues can be IR-isoform-preferential and that the insulin analogues show isoform-selective effects both in vitro and in vivo. This provides a new and unique tool for future studies of the biological role of the two IR isoforms, as well as exciting and promising potentials for improved insulin therapy in the future.
Sara Vienberg, Stephan Bouman, Heidi Sørensen, Carsten Stidsen, Thomas Kjeldsen, Tine Glendorf, Anders Sørensen, Grith Olsen, Birgitte Andersen and Erica Nishimura designed, performed and/or supervised the experiments. Sara Vienberg, Birgitte Andersen and Erica Nishimura wrote the paper. All authors discussed the results and commented on the paper prior to submission.
This work was supported by Novo Nordisk and the Danish Ministry of Science.
Mette Frost, Tina Kisbye, Irena Duch and Lenette Jørgensen provided technical support and assistance for the experiments.
Abbreviations: BHK, baby-hamster kidney; EDL, extensor digitorum longus; FP, forward primer; GLUT, glucose transporter; HI, human insulin; HSA, human serum albumin; IGF, insulin-like growth factor; IR, insulin receptor; LYD, (Landrace/Yorkshire/Duroc; FIRKO, fat-specific IR knockout; KRH, Krebs–Henseleit; NMRI, Naval Medical Research Institute; RP, reverse primer; RT, reverse transcription
- © The Authors Journal compilation © 2011 Biochemical Society