Treatment with 1,25D₃ to non-diabetic rats increased the rat body weights but did not affect any parameter in the plasma of these animals (Table 1). The injection of streptozotocin reduced body weights and induced hyperglycemia and hypoinsulinemia (Table 1). Treatment with 1,25D₃ to streptozotocin-induced diabetic rats increased body weights but did not revert the hyperglycemia and the hypoinsulinemia induced by the diabetes, while calcium and phosphorus plasma levels were increased and plasma levels of 25-hydroxyvitamin D₃ were maintained in the normal range 30 (Table 1).

Treatment with 1,25D₃ to non-diabetic rats did not affect any of the above cited urine parameters under study. Streptozotocin injection induced glycosuria (> 500 mg/dl) and the appearance of leukocytes, ketones and blood in the urine of these animals. Treatment with 1,25D₃to streptozotocin-induced diabetic rats did not modify the glycosuria and the ketonuria of the diabetes, but partially decreased the number of leukocytes in urine (from 10–25 to 5–10 leukocytes/μl).

Treatment with 1,25D₃ to non-diabetic rats increased both the protein content and the indicator of cell size (protein/DNA) in the kidney, liver and adipose tissue of these animals while the DNA and RNA content remained unaltered (Table 2). Streptozotocin injection incremented the protein content in these three tissues, accompanied in the case of kidney with an increment of the protein/DNA ratio and in the case of the liver with an increment of the DNA content. Treatment with 1,25D₃ of streptozotocin-induced diabetic rats practically did not modify neither the high values of protein induced by the diabetes in the three tissues nor the increases in the DNA content in kidney and liver. In addition, this treatment did not alter the RNA content in any of these three tissues (Table 2).

Northern blot assays of kidney (Figure 1, panels A and B), liver (Figure 1, panels C and D) and epididymal adipose tissue (Figure 1, panels E and F) from control animals revealed two major IR mRNA species of approximately 9.5 and 7.5 Kb in size. The relative amounts of the two IR mRNA species measured as 9.5 Kb/7.5 Kb ratio were: 2.2 ± 0.2 in kidney, 1.2 ± 0.2 in liver and 1.1 ± 0.03 in adipose tissue. Treatment with 1,25D₃ to non-diabetic rats did not affect any of the two IR mRNA species in any of the three tissues studied (Figure 1, panels A to F). IR mRNA species were expressed per unit of RNA in view of the observation (Table 2) that treatment with 1,25D₃ to both non-diabetic and diabetic rats did not alter RNA content per gram of tissue in any of the three tissues. The induction of diabetes by streptozotocin produced important increments in the levels of both IR mRNA species in liver and adipose tissue but not in kidney. In particular, 78% increase corresponding to the 9.5 Kb and 69% to the 7.5 Kb species in liver (Figure 1, panel D), and 58% increase corresponding to the 9.5 Kb and 92% increase to the 7.5 Kb species in the adipose tissue (Figure 1, panel F). Treatment with 1,25D₃ to diabetic rats partially prevented the over-expression of the IR mRNA in the liver (Figure 1, panels C and D) and adipose tissue (Figure 1, panels E and F) of these animals. The 9.5 Kb species was particularly decreased in liver and the 7.5 kb species in the adipose tissue. However, treatment with 1,25D₃ to diabetic rats did not affect IR expression in kidney (Figure 1, panels A and B).

Insulin binding studies showed that treatment with 1,25D₃ to non-diabetic rats did not alter IR number or IR affinity as reflected by the ED:50 value in adipocytes of these animals (Figure 2 and Table 3). Streptozotocin injection produced a 64% decrease in the number of IRs without altering IR affinity. Treatment with 1,25D₃ to streptozotocin-induced diabetic rats inverted the decrease in the IR number induced by the diabetes to values not significantly different from those observed in control adipocytes and without altering IR affinity (Figure 2 and Table 3).

Given that both the weight of the adipose tissue and the diameter of the adipocytes were markedly decreased by the diabetes (Table 3), we also calculated the IR number per cell surface area (μm²). In these conditions, although the effect of the diabetes decreasing IR number was somewhat minor (44%), the tendency of 1,25D₃ to normalize the IR number was also evident (Table 3).

Treatment with 1,25D₃ to non-diabetic rats did not alter basal or insulin-stimulated glucose transport in adipocytes from these animals (Figure 3). The injection of streptozotocin clearly decreased both basal and insulin-stimulated glucose transport. The treatment with 1,25D₃ to streptozotocin-induced diabetic rats, improved by 107% the decreased basal glucose transport and by 71% the insulin-stimulated glucose transport in adipocytes from these diabetic animals (Figure 3).

We performed a computer search of virtual VDRE sequences in the rat IR gene promoter by homology with two consensus VDRE sequences. The first consensus (5'GGGTCANNGGGGGCA3') was previously compiled and reported by us from a series of described functional VDRE sequences in various 1,25D₃-stimulated promoters 29. The search with this first consensus indicated no sequence identical to this VDRE sequence, and neither was any sequence found to show 1, 2, or 3 base variations. Nevertheless, we detected three virtual VDRE sequences showing a difference of four bases from this consensus: -249/-235, -637/-623 and -916/-902 (Figure 4, panel A). Comparison of the 3' half-element of each of these sequences with the 3' half-element of this first VDRE consensus, indicated complete homology only in the case of the -916/-902 sequence, and two base variations in the other two (Figure 4, panel A).

The second computer search of virtual VDREs, now by homology with the second consensus VDRE sequence (5'PuGGTCANNPuPuGTTCA3'), was also performed. Our consensus and the latter proposed by Colnot et al. 27, and Haussler et al. 26 have identical 5' half-elements, but ours contains two GG instead of two TT, in the 3' half-element. The spacers are identical, with a G or Pu at position 3, which appears to be important in VDR binding. The results of this computer search by homology with this second consensus indicated no sequence identical to this VDRE, and neither was any sequence found to show 1, 2, or 3 base variations. Nevertheless, we detected three virtual VDRE sequences showing a difference of four bases from this second consensus: -25/-11, -247/-233 and -881/-867 (Figure 4, panel A). Of these, the -247/-233 sequence was in part the same as that of the -249/-235 sequence detected with the first consensus (Figure 4, panel A). However, the 3' half-element of this second sequence has only one base variation instead of two that the first sequence had (Figure 4, panel A). Therefore, in principle there were six virtual VDRE sequences (Figure 4, panel A).

The next step was to check for the possibility of cis-elements related to these virtual VDRE sequences in the rat IR gene promoter. Thus, a computer search of AP-1 and AP-2 sites by homology with the AP-1 site consensus (5'TGAC/GTCA3') 31, and the AP-2 site consensus (5'GCCN3GGG3') 32, respectively, was performed.

Exploring the AP-1 sites, we detected two AP-1-like sites showing a difference of only one base from this consensus: the -374/-368 (5'TGTCTCA3') and the -963/-957 (5'TGAGCCA3'). However, these two sites are neither adjacent nor overlapped in any of our virtual VDRE sequences. Nevertheless, as far as we know, this is the first mention of these AP-1-like sites in the rat IR gene promoter. With respect to the AP-2 sites, we detected seventeen AP-2-like sites showing a difference of only one base from this consensus: -71/-63, -79/-71, -171/-163, -195/-187, -230/-222, -245/-237, -253/-245, -405/-397, -465/-457, -480/-472, -545/-537, -634/-626, -635/-627, -642/-634, -650/-642, -1087/-1079, -1088/-1080. Of these AP-2-like sites, only seven are included in panel B of the Figure 4. These were those flanking or overlapping some of the potential VDRE sequences indicated in panel A of the Figure 4. Of them, the -230/-222 (1) was located downstream, and the -245/-237 (2) and the -253/-245 (3) were overlapping the virtual VDRE sequences: -249/-235 and -247/-233, as shown in Figure 4, panel C. Moreover, the AP-2-like sites: -634/-626 (4) and -635/-627 (5) (Figure 4, panel B) were overlapping the virtual VDRE sequence: -637/-623 (Figure 4, panel C), and the AP-2-like sites: -642/-634 (6) and -650/-642 (7) (Figure 4, panel B) were located in tandem upstream of this last virtual VDRE sequence: -637/-623 (Figure 4, panel C).

Therefore, we postulate as candidate VDRE sequences overlapped by AP-2-like sites those represented in Figure 4, panel C. One, located between -256 and -219 bp with three AP-2-like sites, and the other extending from -653 and -620 bp of the rat IR gene promoter with four AP-2-like sites. Separately or together, these VDRE sequences could form a locus that may respond to 1,25D3 via activation of VDR.

Treatment with 1,25D₃ to non-diabetic rats increased body weights but it did not affect any parameter measured in the plasma or urine of these animals, including the plasma levels of calcium, phosphorus and 25-hydroxyvitamin D₃. Indeed, we did not detect significant differences in calcium and phosphorus, despite the increases of approximately 30%. In relation to 25-hydroxyvitamin D₃ a number of studies 33 has shown that human serum values of 25-hydroxyvitamin D₃ were maintained within a normal range across vitamin D supplies from 80 to 1000 IU daily, indicating the existence of an important homeostatic control system regulating human 25(OH)D₃ serum concentration. Considering this, in the present study we hypothesize the existence of a similar counter regulatory mechanism in the rat. This mechanism could maintain rat plasma values of 25-hydroxyvitamin D₃ within the observed normal range of 15–19 ng/ml, with vitamin D supplies of 150 IU daily administered by injection plus the vitamin D consumed in the diet.

Treatment with 1,25D₃ increased the protein content and the indicator of cell size (protein/DNA) in the kidney, liver, and adipose tissue of non-diabetic rats, although the DNA and RNA content remained unaltered. These increments might indicate hypertrophy. However, although treatment with 1,25D₃ increased rat body weights, it did not particularly augment the weight of these tissues. Another explanation is that 1,25D₃ might cause hyperplasia. Regarding this, conflicting results have been reported 3435.

We also observed that treatment with 1,25D₃ to non-diabetic rats did not affect the expression of any of the two 9.5 and 7.5 Kb IR mRNA species present in kidney, liver, and adipose tissue. Our group and others have previously reported the presence of these species and their relative proportions in rat tissues 2136. Since there is only one rat IR gene, variation in transcript length may reflect alternative RNA processing events, such as polyadenylation at different 3' end sites in the final untranslated exon 37.

In addition, we detected that treatment with 1,25D₃ to non-diabetic rats affected neither the IR number nor the insulin response in terms of glucose transport in isolated adipocytes from these animals.

The lack of effects of 1,25D₃ on IR expression and insulin activity in tissues of non-diabetic rats could be related to the absence of VDR in these tissues. However, both the VDR protein and VDR mRNA have been detected in the rat kidney 3839 and liver 4041, and while a VDR-like protein was identified in 3T3-L1 adipose cells 42 VDR mRNA has been found in rat perirenal adipose tissue and 3T3-L1 adipose cells 43. Another possibility could be the absence of regulation of VDR by its ligand in these tissues. Such lack of regulation has been described in the kidney 39. In the liver there are no previous data, while in adipose cells treatment with 1,25D₃ led to an increase in VDR mRNA levels 43 and a decrease in VDR protein 35. Regarding this and given our previous experience in determining both VDR protein and VDR mRNA expression and their regulation by 1,25D₃ in human cells 67, we determined the expression of this receptor and its regulation by 1,25D₃ at the protein level in the kidney, liver and adipose tissue of non-diabetic rats. Our results showed a VDR protein only in the kidney (data not shown). More experiments are necessary to confirm this.

Treatment with 1,25D₃ to streptozotocin-induced diabetic rats increased their body weight, but it did not revert the hyperglycemia and the hypoinsulinemia provoked by the streptozotocin. We also detected normal levels of 25-hydroxyvitamin D₃ in streptozotocin-induced diabetic rats in accordance with other authors 30. These unchanged levels of 25-hydroxyvitamin D₃ in streptozotocin-induced diabetic rats neither were altered by the treatment with 1,25D₃, possibly due to the postulated plasma homeostatic control system that could regulate 25-hydroxyvitamin D₃ plasma levels also in the absence of insulin in both groups of diabetic rats. In addition, we observed increased plasma levels of calcium and phosphorus after the treatment with 1,25D₃. With regard to these last parameters, it is known that calcium per se is important for insulin secretion, as well as for correction of glucose intolerance 4445. Moreover, calcium and phosphorus have been described as regulators of VDR in renal and intestinal tissues 3946.

Treatment with 1,25D₃ to streptozotocin-diabetic rats did not modify the glycosuria and the ketonuria induced by the diabetes, but partially decreased the number of leukocytes in urine. This latter effect could be due to the broadly reported anti-inflammatory effects of 1,25D₃ and/or to a possible anti-proliferative effect of this hormone, as observed by our group and others in cell culture 633.

With respect to tissue effects of 1,25D₃ in streptozotocin-induced diabetic rats, we observed the permanence of high protein and DNA content values induced by the diabetes, accompanied by slight increases of the protein/DNA ratio values in the kidney and liver, but not in the adipose tissue. This suggests a certain hypertrophy produced by the action of 1,25D₃ in these two tissues. However, although treatment with 1,25D₃ to streptozotocin-diabetic rats increased rat body weights, it did not increase the weight of the kidney, liver or adipose tissue. The other explanation, hyperplasia, has conflicting results 3334.

The induction of diabetes by streptozotocin did not affect IR mRNA species in the kidney. This finding was not in accord with previous results of increased renal IR mRNA levels, quantified by slot blot, in a similar rat model 22. This may reflect the failure to alter VDR number in the kidney of streptozotocin-diabetic rats 34. The treatment with 1,25D₃ to streptozotocin-induced diabetic rats neither affected IR mRNA species in the kidney.

Contrarily, streptozotocin produced an important increment in the levels of both IR mRNA species in liver. These findings agree with previous results of increased IR mRNA levels quantified by slot blot 22, Northern blot analysis 2147 and increased IR gene transcription 21 in the liver of streptozotocin-induced diabetic rats. Streptozotocin also increased the levels of both IR mRNA species in the adipose tissue. This increment represents the first demonstration of an in vivo regulation of IR mRNA levels in the adipose tissue of streptozotocin-induced diabetic rats.

In addition, we observed that the treatment with 1,25D₃ to streptozotocin-induced diabetic rats partially prevented the over-expression of IR mRNA induced by the diabetes in the liver and adipose tissue of these animals. The 9.5 Kb IR mRNA species was particularly decreased by 1,25D₃ in the liver, and the 7.5 Kb species in the adipose tissue. Therefore, these results provide the first demonstration of an in vivo tissue-specific regulation of rat IR mRNA levels by 1,25D₃ in streptozotocin-induced diabetic rats.

A role for VDR in such genomic actions of 1,25D₃ in liver and adipose tissue is possible, although to our knowledge the regulation of VDR by 1,25D₃ has not yet been studied in these tissues of streptozotocin-diabetic rats.

The correction by 1,25D₃ of the over-expression of IR expression at the RNA level in the liver and adipose tissue of streptozotocin-diabetic animals could represent a beneficial fact of amelioration of the diabetes. In this sense, other facts of amelioration of diabetes were the almost normalization of the number of IRs in adipocytes of streptozotocin-diabetic animals treated with 1,25D₃, without alterations in receptor affinity, and the improvement of both basal and insulin-stimulated glucose transport in adipocytes of these diabetic animals treated with 1,25D₃.

The present in vivo results do not allow to determine whether the prevention by 1,25D₃ of the over-expression of IR mRNA induced by the diabetes in liver and adipose tissue of streptozotocin-diabetic animals is due to transcriptional and/or post-transcriptional regulation. Nevertheless, our previous in vitro results demonstrated that 1,25D₃ increased IR mRNA levels via mechanisms involving direct transcriptional activation of the human IR gene 567. Moreover, we identified an active VDRE in the human IR gene promoter that accounted for the transcriptional induction of this gene by 1,25D₃ in U-937 cells 29.

Therefore, in support of the possible participation of a transcriptional regulation by 1,25D₃ in the present in vivo processes, we have detected by computer search in the rat IR gene promoter two candidate VDREs overlapped by various AP-2-like sites. One VDRE is located between -256 and -219 bp, while the other extends from -653 and -620 bp of the rat IR gene promoter (Figure 4, panel C). Individually or in conjunction, these potential VDREs could form a locus that may respond to 1,25D₃ via activation of VDR. This locus may mediate the cross-talk between vitamin D and insulin signalling, although this remains to be determined.

Four groups of male Wistar rats weighing 200–220 g at the onset were used in this study. They were kept under standard conditions of light and temperature and given free access to tap water and standard laboratory chow (Panlab, A04) containing 8.8 mg/g of calcium, 5.9 mg/g of phosphorus and 1.5 IU/g of vitamin D. The first group comprised of non-diabetic rats receiving sham-treatments during the 15 days of the experimental period (Control-rats). The second group included non-diabetic rats i.p. injected with 1,25D₃ (Calcijex, Abbot) [150 IU/Kg (3.75 μg/Kg) one a day, for 15 days] (1,25D₃-rats). The third group consisted of rats rendered diabetic by a single injection of streptozotocin (Sigma) [65 mg/kg] on the 8th day of the 15-day period (STZ-rats). Finally, the fourth group comprised of rats rendered diabetic by streptozotocin on the 8th day, and i.p. injected with 1,25D₃ [150 IU/Kg (3.75 μg/Kg) one a day, for 15 days] (STZ+1,25D₃-rats). The National Guide for the Care and Use of Laboratory Animals was strictly followed during this study.

Urine samples were collected before decapitation without anaesthesia and trunk blood samples were recovered for plasma measurements. On sacrifice, the kidney, liver and epididymal adipose tissue were immediately removed, individually packed and frozen in liquid nitrogen for nucleic acid and protein determinations. Isolated adipocytes were obtained from fresh epididymal adipose tissue for insulin binding and insulin activity assays.

Different parameters in urine were measured using the Combur Test (Roche). Plasma insulin levels were determined by radioimmunoassay using rat insulin as standard. Plasma concentrations of glucose, 25-hydroxyvitamin D₃, calcium, phosphorus and proteins were estimated using commercially available techniques. The total DNA and RNA contents of the homogenates from individual tissues were estimated by spectrofluorometry. The protein content was determined by the Bradford method.

For Northern blot assays, RNA samples (40 μg) of kidney, liver and epididymal adipose tissue, were denatured, electrophoresed in 1.1% agarose-formaldehyde gels and blotted onto nylon membranes 36. Ethidium bromide staining of the 28S and 18S ribosomal RNAs was routinely checked before blotting as a control of sample loading and after blotting as a control of RNA transfer. RNA blots were prehybridized, hybridized with excess [³²P]-labelled probe (the 0.98-Kb rat IR specific EcoR1 fragment of the p16 clone, gift from Prof. Goldstein), washed under stringent conditions and finally autoradiographed, as described previously 36. The autoradiographs were scanned with a laser densitometer and the readings normalized with the respective amounts of 28S rRNA, as revealed by ethidium bromide.

Binding assays were carried out as described previously 48. In brief, isolated adipocytes (0.25 × 10⁶ cells/ml) were incubated at 30°C for 30 min in 400 μl Krebs-Hepes buffer pH 7.6 with mono-[¹²⁵I]-insulin (0.2 × 10^-9 M) (NEN Life Sciences), in the presence or absence of increasing concentrations (0.2 × 10^-10 to 0.5 × 10^-7 M) of unlabelled insulin (Sigma). Non-specific binding was determined in the presence of 1 × 10^-7 M unlabelled insulin and subtracted from each point. The adipocyte number was determined with a Neubauer-type hemocytometer and cell viability (assessed by 0.2% trypan blue exclusion) was greater than 90%. Measurements of the cell diameter were also performed. The data were analysed by the Scatchard method using the LIGAND program of Munson and Rodbard 49.

Glucose transport was measured in isolated adipocytes by determining the uptake of D-[¹⁴C(U)]-glucose (NEN Life Sciences) at trace concentrations using our modification of the method of Kashiwagi et al. 50. This method is based on the premise that glucose uptake provides a measure of glucose transport when studies are carried out at very low glucose concentrations. In brief, 0.25 × 10⁶ adipocytes/ml were incubated with 5 × 10^-7 M D-[¹⁴C(U)]-glucose for 1 h at 37°C in the presence and absence of a concentration of insulin (10^-8 M) that gave maximal glucose transport in earlier experiments 6. Assays were stopped by washing the cells with ice-cold PBS. The cells were solubilized in 0.5% SDS and 0.1 N NaOH and the associated radioactivity was counted.

The identification of virtual VDRE sequences in the Rattus norvegicus partial IR promoter [GenBank:AJ006071] was carried out by homology with two consensus VDRE sequences using the SEQFIND programme developed in our laboratory and previously employed in the computer analysis of various DNA sequences including VDREs 29. The first consensus (5'GGGTCANNGGGGGCA3') had been compiled and reported by us from a series of described functional VDRE sequences in various 1,25D₃-stimulated promoters 29. The second consensus VDRE sequence (5'PuGGTCANNPuPuGTTCA3') had been proposed by Colnot et al. 27, and Haussler et al. 26. The presence of AP-1 and AP-2 sites flanking or overlapping the potential VDREs were identified by their homology with the AP-1 site consensus (5'TGAC/GTCA3') 31 and the AP-2 site consensus (5'GCCN3GGG3') 32 respectively, also using the SEQFIND programme.

Unless otherwise indicated, the data are expressed as the mean ± SEM. A comparison between the groups was carried out using the two-tailed, unpaired Student t-test and/or ANOVA comparison, as appropriate. Differences were considered statistically significant when p < 0.05.