Net synthesis of pancreatic β-cells peaks before 2 years of life. β-Cell mass is set within the first 5 years of life. In-frame translational readthrough of the NRP1 gene exon 9 into intron 9 generates a truncated neuropilin-1 protein lacking downstream sequence necessary for binding VEGF that stimulates β-cell replication. VEGF is critical for developing but not adult islet neogenesis. Herein we show that cells in human pancreatic islets containing the full-length neuropilin-1 possess insulin but cells that contain the truncated neuropilin-1 are devoid of insulin. Decreased insulin cells increases susceptibility to onset of type 1 diabetes at a younger age. We also show that the frequency of a genetic marker in NRP1 intron 9 is higher among patients with onset of type 1 diabetes before age 4 years (31.8%), including those with onset at 0.67–2.00 and 2–4 years, compared with that in patients with onset at 4–8 years, at 8–12 years, and after 16 years (16.1%) with frequency equal to that in subjects without diabetes (16.0%). Decreased insulin cells plus the genetic data are consistent with a low effect mechanism that alters the onset of type 1 diabetes to a very young age in some patients, thus supporting the endotype concept that type 1 diabetes is a heterogeneous disease.

We previously used denaturing high-performance liquid chromatography to screen a small number of pediatric patients with type 1 diabetes for 21 candidate genes in a region of chromosome 10 suspected to contain (a) susceptibility genes(s) for type 1 diabetes. The only association we found was with two minor alleles in intron 9 of the NRP1 gene (1). NRP1 encodes neuuropilin-1, a transmembrane receptor expressed on the surface and in the cytosol of cells. The data hinted that the associations might be stronger with diabetes onset prior to age 10 years. Recently, another report mentioned associations of the NRP1 gene with type 1 diabetes, but the report did not mention the patients’ ages of diabetes onset (2). We discovered that in the human NRP1 gene, intron 9 is in frame with exon 9, allowing translational readthrough from exon 9 into intron 9. This readthrough produces an abnormal truncated nonfunctional neuropilin-1 protein encoded by NRP1 exons 1–9 plus a 59–amino acid segment encoded by NRP1 intron 9 that is terminated by a stop codon. Although this abnormal neuropilin-1 isoform was initially known only in humans (1), more recent data from the National Center for Biotechnology Information database corroborate that this isoform is also present in some nonhuman primates. No nonprimates possess this isoform (Supplementary Table 1).

Neuropilin-1 is a coreceptor for vascular endothelial growth factor (VEGF). There is extensive evidence that VEGF is expressed in human pancreatic β-cells and is essential for their neogenesis and proliferation apart from vascularization (3–7). The b1b2 domain of neuropilin-1 that is encoded by exons 6 through 10 of NRP1 enables neuropilin-1 to bind to VEGF (8). Any isoform of neuropilin-1 lacking any part of the b1b2 domain prevents VEGF binding, and this will prevent VEGF from stimulating β-cell replication (9). In humans VEGF is critical for developing but not adult pancreatic islet neogenesis (10). Also in humans, the net synthesis of β-cells peaks before 2 years of life and β-cell mass is set within the first 5 years of life (11).

Herein we show that in human pancreas neuropilin-1 is expressed only in the islets (Supplementary Fig. 1). We also show that in human pancreas the islet cells that possess the normal whole-length neuropilin-1 isoform contain insulin but the cells in the islets that possess the NRP1 intron 9–encoded truncated protein sequence are devoid of insulin. This can be explained by the stop codon in intron 9 of the NRP1 gene resulting in the synthesis of the truncated neuropilin-1 isoform that lacks the b2 domain protein sequence encoded by NRP1 exon 10. This isoform cannot bind VEGF that stimulates normal β-cell replication. There are also well-described soluble isoforms of neuropilin-1 that lack the sequence within the b2 domain that antagonize normal neuropilin-1–mediated cellular activities (8,12). Either mechanism will decrease the formation of normal β-cells from the beginning of life and increase the susceptibility to type 1 diabetes at a young age.

We also performed a genetic study of patients with type 1 diabetes from our center that, unlike with most previous studies, included a large percentage of patients with onset of diabetes before age 18 years. Consistent with the mechanism caused by the truncated neuropilin-1 isoform decreasing the percentage of β-cells early in life, we found that a genetic marker in NRP1 intron 9 is associated with a low effect that shifts the onset of type 1 diabetes to a very young age (<2 and <4 years). The mechanism of the truncated isoform of neuropilin-1 decreasing the percentage of β-cells in pancreatic islets from the beginning of life coupled with the genetic data supports the idea that type 1 diabetes is a heterogeneous disease.

Analyses

Histochemistry analysis of human pancreas (1), measurement of mRNA in human pancreatic islets (13), sequencing of DNA in peripheral blood buffy coat cells (14), and HLA DR typing (15) were performed with standard methods as previously described.

Subjects

Medical histories and blood samples were obtained from patients attending the University of Wisconsin School of Medicine and Public Health Pediatric Diabetes Clinic or their relatives with diabetes according to a protocol approved by the University of Wisconsin Health Sciences Institutional Review Board. All patients with type 1 diabetes and non–type 1 diabetes control subjects studied were Caucasian because > 97% of the patients attending the pediatric diabetes clinic were Caucasian. All patients with onset of diabetes before age 18 years, as well as many of their parents with diabetes onset when they were children, were patients of M.J.M. (91% of the total patients analyzed). More information on patient selection can be found in Supplementary Material.

Statistical Analysis

Rates of NRP1 alleles were calculated for each age-group and reported along with the corresponding 95% CIs. Logistic regression analyses were conducted to compare the rates between age-groups. The Šidák step-down method for multiple comparisons was used in conducting pairwise comparisons between groups. Linear regression analysis and the Cochran-Armitage trend test were conducted to evaluate the association between age and the rates. Percentages of HLA and NRP1 alleles were summarized in tabular format. All reported P values are two sided, and P < 0.05 was used to define statistical significance. Statistical analyses were conducted with SAS software (SAS Institute, Cary, NC), version 9.4.

Data and Resources

Data and resources are available on request.

Histochemistry Analysis Shows a Truncated Neuropilin-1 Isoform Encoded by NRP1 Intron 9 Is Unable to Stimulate Insulin Synthesis

Supplementary Fig. 1 is a low-power micrograph of human pancreas that shows that antibodies we raised against three different amino acid sequences of neuropilin-1 react only with pancreatic islets. Antibody 30 was raised against a neuropilin-1 amino acid sequence encoded by the NRP1 gene exon 4 and antibodies 23 and 2733 were raised against two different neuroplin-1 amino acid sequences encoded by the NRP1 gene intron 9. It was previously shown that none of the three antibodies stain the glucagon-containing α-cells in the islets when the antibodies are adsorbed with polylysine to eliminate nonspecific binding to glucagon (1).

Since intron 9 of the NRP1 gene possesses a stop codon, any transcriptional readthrough from the exon 9 gene sequence into the intron 9 gene sequence will synthesize a truncated neuropilin-1 isoform terminated after the protein sequence encoded by NRP1 exons 1 through 9 plus the 59 amino acid sequence encoded by the intron 9 gene sequence up to the stop codon. Since NRP1 exon 4–encoded amino acid sequence is synthesized in front of exon 9 and/or intron 9 amino acid sequence, any amino acid sequence encoded by exon 4 will be present in both the normal full-length neuropilin-1 isoform containing the NRP1 exon 9–encoded amino acid sequence alone and the truncated neuropilin-1 isoform encoded by transcriptional readthrough of NRP1 exon 9 into intron 9. Thus, antibody 30 will stain the normal full-length neuropilin-1 protein sequence as well as the truncated neuroplin-1 protein. Antibodies 23 and 2733 will stain only the truncated neuropilin-1 isoforms.

We stained pancreases from two different humans without diabetes with an anti-insulin antibody and also with either antibody 30 or antibody 23 (Fig. 1). Antibody 23 (green color) did not stain the same cells that the anti-insulin antibody stained (Ins, red color), as shown in the panels where the antibodies were applied to separate pancreas sections or when both antibodies were applied to the same section. In contrast, when pancreas sections were stained with both antibody 30 (green color when stained with antibody 30 alone) plus anti-insulin antibody (red color when stained with anti-insulin alone) there were cells in which the two colors merged into an orange-yellow color, indicating that these cells contained both insulin and the full-length neuropilin-1 isoform. However, in these same panels of the pancreas stained with both the anti-insulin antibody and antibody 30, there were also other cells that remained green in color, indicating that they did not contain insulin but contained the abnormal truncated isoform of neuropilin-1 encoded by NRP1 intron 9. These results show that the cells that contained the truncated NRP1 intron 9–encoded neuropilin-1 protein isoform that lacks amino acid sequence encoded by NRP1 exon 10 and higher number exons were unable to bind to VEGF that is a strong stimulator of insulin cell neogenesis (3–9) especially in early life in humans (10,11).

Figure 1
In human pancreatic islets, cells with NRP1 intron 9–encoded amino acid sequence are devoid of insulin. Consecutively cut sections of one pancreas (top two rows of panels) and a second human pancreas (bottom two rows of panels) were stained with antibody 23 that reacts with NRP1 intron 9–encoded amino acid sequence alone (green stain) or against antibody 30 that reacts with NRP1 exon 4–encoded amino acid sequence alone (green stain) and/or with anti-insulin antibody alone (red stain). Only pancreatic islet cells took up the stains. The sections that were first stained with antibody 23 alone or antibody 30 alone were then also stained with anti-insulin antibody. In the pictures where the pancreases were stained with antibody 30 plus anti-insulin antibody, there were cells that stained an orange-yellow color. This indicates that these cells contained the full-length neuropilin-1 protein sequence plus insulin. In these same pictures of pancreases that were stained with both of these antibodies there were cells that remained green. This indicates that these green-colored cells did not contain insulin. In pancreas sections stained with anti-insulin antibody plus antibody 23, which reacts with NRP1 intron 9–encoded amino acid sequence, the cells were strained either green or red and there was no merging of the two colors into an orange-yellow color. Thus, all cells with intron 9–encoded amino acid sequence were devoid of insulin. The results are consistent with the isoform of neuropilin-1 that is truncated by the stop codon in NRP1 intron 9 not containing NRP1 exon 10 and higher number exon–encoded amino acid sequence required for binding to VEGF not being able to stimulate insulin synthesis. Orange-yellow coloring in the merged pictures for cells stained with antibody 30 and anti-insulin antibody indicates that they contained the full-length neuropilin-1 that was able to bind to VEGF and stimulate insulin synthesis. Pancreases were selected that had islets with relatively high amounts of NRP1 intron 9–encoded amino acid sequence to better illustrate the differences in cell types.
View largeDownload slide

In human pancreatic islets, cells with NRP1 intron 9–encoded amino acid sequence are devoid of insulin. Consecutively cut sections of one pancreas (top two rows of panels) and a second human pancreas (bottom two rows of panels) were stained with antibody 23 that reacts with NRP1 intron 9–encoded amino acid sequence alone (green stain) or against antibody 30 that reacts with NRP1 exon 4–encoded amino acid sequence alone (green stain) and/or with anti-insulin antibody alone (red stain). Only pancreatic islet cells took up the stains. The sections that were first stained with antibody 23 alone or antibody 30 alone were then also stained with anti-insulin antibody. In the pictures where the pancreases were stained with antibody 30 plus anti-insulin antibody, there were cells that stained an orange-yellow color. This indicates that these cells contained the full-length neuropilin-1 protein sequence plus insulin. In these same pictures of pancreases that were stained with both of these antibodies there were cells that remained green. This indicates that these green-colored cells did not contain insulin. In pancreas sections stained with anti-insulin antibody plus antibody 23, which reacts with NRP1 intron 9–encoded amino acid sequence, the cells were strained either green or red and there was no merging of the two colors into an orange-yellow color. Thus, all cells with intron 9–encoded amino acid sequence were devoid of insulin. The results are consistent with the isoform of neuropilin-1 that is truncated by the stop codon in NRP1 intron 9 not containing NRP1 exon 10 and higher number exon–encoded amino acid sequence required for binding to VEGF not being able to stimulate insulin synthesis. Orange-yellow coloring in the merged pictures for cells stained with antibody 30 and anti-insulin antibody indicates that they contained the full-length neuropilin-1 that was able to bind to VEGF and stimulate insulin synthesis. Pancreases were selected that had islets with relatively high amounts of NRP1 intron 9–encoded amino acid sequence to better illustrate the differences in cell types.

Figure 1
In human pancreatic islets, cells with NRP1 intron 9–encoded amino acid sequence are devoid of insulin. Consecutively cut sections of one pancreas (top two rows of panels) and a second human pancreas (bottom two rows of panels) were stained with antibody 23 that reacts with NRP1 intron 9–encoded amino acid sequence alone (green stain) or against antibody 30 that reacts with NRP1 exon 4–encoded amino acid sequence alone (green stain) and/or with anti-insulin antibody alone (red stain). Only pancreatic islet cells took up the stains. The sections that were first stained with antibody 23 alone or antibody 30 alone were then also stained with anti-insulin antibody. In the pictures where the pancreases were stained with antibody 30 plus anti-insulin antibody, there were cells that stained an orange-yellow color. This indicates that these cells contained the full-length neuropilin-1 protein sequence plus insulin. In these same pictures of pancreases that were stained with both of these antibodies there were cells that remained green. This indicates that these green-colored cells did not contain insulin. In pancreas sections stained with anti-insulin antibody plus antibody 23, which reacts with NRP1 intron 9–encoded amino acid sequence, the cells were strained either green or red and there was no merging of the two colors into an orange-yellow color. Thus, all cells with intron 9–encoded amino acid sequence were devoid of insulin. The results are consistent with the isoform of neuropilin-1 that is truncated by the stop codon in NRP1 intron 9 not containing NRP1 exon 10 and higher number exon–encoded amino acid sequence required for binding to VEGF not being able to stimulate insulin synthesis. Orange-yellow coloring in the merged pictures for cells stained with antibody 30 and anti-insulin antibody indicates that they contained the full-length neuropilin-1 that was able to bind to VEGF and stimulate insulin synthesis. Pancreases were selected that had islets with relatively high amounts of NRP1 intron 9–encoded amino acid sequence to better illustrate the differences in cell types.
View largeDownload slide

In human pancreatic islets, cells with NRP1 intron 9–encoded amino acid sequence are devoid of insulin. Consecutively cut sections of one pancreas (top two rows of panels) and a second human pancreas (bottom two rows of panels) were stained with antibody 23 that reacts with NRP1 intron 9–encoded amino acid sequence alone (green stain) or against antibody 30 that reacts with NRP1 exon 4–encoded amino acid sequence alone (green stain) and/or with anti-insulin antibody alone (red stain). Only pancreatic islet cells took up the stains. The sections that were first stained with antibody 23 alone or antibody 30 alone were then also stained with anti-insulin antibody. In the pictures where the pancreases were stained with antibody 30 plus anti-insulin antibody, there were cells that stained an orange-yellow color. This indicates that these cells contained the full-length neuropilin-1 protein sequence plus insulin. In these same pictures of pancreases that were stained with both of these antibodies there were cells that remained green. This indicates that these green-colored cells did not contain insulin. In pancreas sections stained with anti-insulin antibody plus antibody 23, which reacts with NRP1 intron 9–encoded amino acid sequence, the cells were strained either green or red and there was no merging of the two colors into an orange-yellow color. Thus, all cells with intron 9–encoded amino acid sequence were devoid of insulin. The results are consistent with the isoform of neuropilin-1 that is truncated by the stop codon in NRP1 intron 9 not containing NRP1 exon 10 and higher number exon–encoded amino acid sequence required for binding to VEGF not being able to stimulate insulin synthesis. Orange-yellow coloring in the merged pictures for cells stained with antibody 30 and anti-insulin antibody indicates that they contained the full-length neuropilin-1 that was able to bind to VEGF and stimulate insulin synthesis. Pancreases were selected that had islets with relatively high amounts of NRP1 intron 9–encoded amino acid sequence to better illustrate the differences in cell types.

Close modal

The Expression of NRP1 Intron 9–Encoded mRNA and Protein Varies Within Human Pancreatic Islets

Table 1 shows that among pancreatic islets isolated from 10 human donors the expression of mRNA encoded by readthrough of NRP1 exon 9 into intron 9 was quite variable, ranging from 0 to 39% of the level of mRNA encoded by exon 9. Histochemical quantification of protein isoforms in 21 pancreatic islets present in five human pancreases showed that, similar to the range of NRP1 intron 9–encoded mRNA expression, cells that contained the NRP1 intron 9–encoded truncated neuropilin-1 protein (which do not contain insulin) ranged from 0 to 46% and cells that contained the full-length neuropilin-1 protein (which do contain insulin) ranged from 54 to 100% of the cells that contained a neuropilin-1 isoform (Supplementary Table 2). Beginning life with a low amount of insulin cells in a person’s pancreas will increase the probability of the person developing diabetes at a young age if the person has other risk factors for type 1 diabetes.

Table 1

Variation in relative NRP1 intron 9–encoded mRNA levels in human pancreatic islets among various individuals

SampleSexAge (years)Relative mRNA level (%)
Exon 9–exon 9Exon 9–intron 9 AExon 9–intron 9 B
1 Male 52 100 2 3 
2 Female 55 100 7 7 
3 Male 40 100 4 8 
4 Male 55 100 12 35 
5 Female 38 100 38 20 
6 Male 43 100 8 5 
7 Male 54 100 1 5 
8 Male 30 100 0 0 
9 Male 32 100 2 19 
10 Male 30 100 39 27 
SampleSexAge (years)Relative mRNA level (%)
Exon 9–exon 9Exon 9–intron 9 AExon 9–intron 9 B
1 Male 52 100 2 3 
2 Female 55 100 7 7 
3 Male 40 100 4 8 
4 Male 55 100 12 35 
5 Female 38 100 38 20 
6 Male 43 100 8 5 
7 Male 54 100 1 5 
8 Male 30 100 0 0 
9 Male 32 100 2 19 
10 Male 30 100 39 27 

Quantitative PCR was used to measure the expression of NRP1 intron 9–encoded mRNA levels resulting from readthrough from exon 9 into intron 9 in pancreatic islets from 10 separate human islet donors without diabetes. Two different sets of primers were used to compare the exon 9–intron 9 mRNA levels with the exon 9–encoded mRNA levels, which were defined as 100%.

View Large

A Marker in NRP1 Intron 9 Is Associated With Very Early Onset of Type 1 Diabetes in Children

Among 1,004 patients with type 1 diabetes (probands, parents, and/or siblings) the 31.8% incidence of a single nucleotide polymorphism minor allele (rs2070303) in NRP1 intron 9 in patients with onset of diabetes before age 4 years was twofold higher than in patients with onset after 16 years (16.1%, P = 0.007) and also higher in patients with onset between ages 4 and 8 years (20.8%, P = 0.007) and 8 and 12 years (22.6%, P = 0.028) and in control subjects without type 1 diabetes (16.0%, P = 0.0002) (Table 2). Among patients with diabetes onset after age 16 years (16.1%) the minor allele was not different than among the control subjects. The frequency of the minor allele among all of the patients with type 1 diabetes (24.0%) (that includes the frequency among patients with diabetes onset before age 4 years) was higher than among control subjects who did not have type 1 diabetes (P = 0.011). The Cochran-Armitage trend test showed that there was a significant positive trend in the increasing frequencies of the minor NRP1 allele with decreasing age of onset of diabetes across the five 4-year age intervals spanning diabetes onset at <4 years up to 20 years plus >20 years (P < 0.03) or the eleven 4-year age intervals spanning diabetes onset at <4 years up to 44 years plus >44 years (P = 0.02).

Table 2

Increased prevalence of the minor allele (T) of NRP1 SNP rs2070303 among children with onset of type 1 diabetes before age 4 years

NCases of T/C + T/TFrequency, %Lower 95% CIUpper 95% CIPP†
Age of onset of type 1 diabetes (years)        
 <4 192 61 31.8 25.6 38.7   
 4–8 279 58 20.8 16.4 26.0   
 8–12 266 60 22.6 17.9 28.0   
 12–16 180 48 26.7 20.7 33.6   
 >16 87 14 16.1 9.8 25.4   
All type 1 diabetes 1,004 241 24.0 21.5 26.8   
Non–type 1 diabetes 231 37 16.0 11.8 21.3   
Comparisons between ages (years) of onset of type 1 diabetes        
 <4 vs. 4–8      0.007 0.043 
 <4 vs. 8–12      0.028 0.082 
 <4 vs. 12–16      0.280 0.482 
 <4 vs. >16      0.007 0.043 
 <4 vs. non–type 1 diabetes      0.0002 0.001 
 >16 vs. non–type 1 diabetes      0.987 0.987 
All type 1 diabetes vs. non–type 1 diabetes      0.011 0.044 
NCases of T/C + T/TFrequency, %Lower 95% CIUpper 95% CIPP†
Age of onset of type 1 diabetes (years)        
 <4 192 61 31.8 25.6 38.7   
 4–8 279 58 20.8 16.4 26.0   
 8–12 266 60 22.6 17.9 28.0   
 12–16 180 48 26.7 20.7 33.6   
 >16 87 14 16.1 9.8 25.4   
All type 1 diabetes 1,004 241 24.0 21.5 26.8   
Non–type 1 diabetes 231 37 16.0 11.8 21.3   
Comparisons between ages (years) of onset of type 1 diabetes        
 <4 vs. 4–8      0.007 0.043 
 <4 vs. 8–12      0.028 0.082 
 <4 vs. 12–16      0.280 0.482 
 <4 vs. >16      0.007 0.043 
 <4 vs. non–type 1 diabetes      0.0002 0.001 
 >16 vs. non–type 1 diabetes      0.987 0.987 
All type 1 diabetes vs. non–type 1 diabetes      0.011 0.044 
*

P value.

†

Adjusted P value for multiple comparisons with Šidák step-down procedure.

View Large

Not shown in Table 2 is that the percentage of patients with the minor NRP1 allele among patients with onset of diabetes before age 4 years (31.8%) was higher (P = 0.02) than among all the patients with type 1 diabetes (24.0%) (that included the patients with onset of type 1 diabetes prior to age 4 years). The minor allele among patients with onset of diabetes between ages 12 and 16 years (26.7%) was higher than among patients with onset after age 16 years (P = 0.05) or control subjects (P = 0.0088).

Separate analyses of the frequencies of the minor allele among age-groups in only the 844 unrelated patients (i.e., probands) from within the total group of 1,004 related and unrelated patients showed that the percentages were very similar to the percentages among the 1,004 patients. The frequency of the minor allele among patients with onset of type 1 diabetes before 4 years of age (33%) was significantly higher than among patients with onset between 4 and 8 years of age (21.5%, P = 0.008) and 8 and 12 years of age (22.7%, P = 0.02) or control subjects without type 1 diabetes (P < 0.0001). Also similar to the 1,004 patients, the frequency of the minor allele among patients with onset of diabetes after 16 years (16.7%) was not significantly different than among the patients without type 1 diabetes (16.0%). (More analyses of the minor allele in the 844 patients are described in Supplementary Material.)

As expected, the frequencies of the minor allele with onset of diabetes between 0.67 and 2.00 years and between 2.00 and 4.00 years among the 844 unrelated patients as well as in the total 1,004 unrelated plus related patients were significantly higher than among patients with onset at older ages. (Details can be found in Supplementary Material.)

The Minor Allele in NRP1 Intron 9 Is Associated With Increased Probability of Developing Diabetes at a Very Young Age Even in Patients With High-Risk HLA DR Alleles

Among the 1,004 related plus unrelated patients with type 1 diabetes, those with high-risk HLA genotypes (HLA DR4 + DR17 and either HLA DR4 or HLA DR17 without the other HLA DR allele) who possessed the minor NRP1 allele had the highest probability (P = 0.000–0.003) of developing diabetes before age 4 years compared with the patients with the same HLA DR genotypes who did not possess the minor NRP1 allele (Table 3 and Supplementary Table 3). This was also true among the 844 unrelated patients with type 1 diabetes (P = 0.000 to 0.004). This indicates that the mechanism associated with the NRP1 intron 9–encoded neuropilin-1 isoform acts independent of the high-risk type 1 diabetes HLA alleles.

Table 3

Higher probability for onset of diabetes before age 4 years in patients who possess the minor allele (T) of NRP1 SNP rs2070303 even when they possess the high-risk type 1 diabetes HLA DR alleles 4, 17, or 4 + 17

HLA typeAge of type 1 diabetes onset (years)Odds ratio for type 1 diabetes (minor allele vs. wild type)Lower 95% CIUpper 95% CIP
HLA ≠ 4, 17,* 15† <4 2.0 0.5 7.8 0.397 
 4–8 1.1 0.4 3.4 0.773 
 8–12 2.2 0.7 6.6 0.177 
 12–16 1.7 0.3 9.0 0.620 
 >16 1.3 0.1 12.1 0.588 
HLA = 4 <4 2.2 1.3 3.7 0.003 
 4–8 1.3 0.8 2.1 0.433 
 8–12 1.7 1.0 2.8 0.052 
 12–16 1.9 1.1 3.2 0.019 
 >16 0.9 0.4 2.0 1.000 
HLA = 17* <4 3.0 1.8 5.0 0.000 
 4–8 1.5 0.9 2.5 0.176 
 8–12 1.4 0.8 2.4 0.223 
 12–16 2.2 1.2 3.8 0.009 
 >16 0.9 0.4 2.4 1.0000 
HLA = 4 + 17* <4 2.7 1.5 5.1 0.002 
 4–8 1.5 0.8 2.8 0.239 
 8–12 1.6 0.8 3.1 0.154 
 12–16 2.2 1.1 4.2 0.033 
 >16 0.6 0.1 2.8 0.746 
HLA = 15† <4 1.7 0.2 17.3 0.509 
 4–8 0.5 0.1 4.2 1.000 
 8–12 1.0 0.1 9.2 1.000 
 12–16 2.6 0.5 14.8 0.257 
 >16 3.5 0.6 21.6 0.192 
HLA typeAge of type 1 diabetes onset (years)Odds ratio for type 1 diabetes (minor allele vs. wild type)Lower 95% CIUpper 95% CIP
HLA ≠ 4, 17,* 15† <4 2.0 0.5 7.8 0.397 
 4–8 1.1 0.4 3.4 0.773 
 8–12 2.2 0.7 6.6 0.177 
 12–16 1.7 0.3 9.0 0.620 
 >16 1.3 0.1 12.1 0.588 
HLA = 4 <4 2.2 1.3 3.7 0.003 
 4–8 1.3 0.8 2.1 0.433 
 8–12 1.7 1.0 2.8 0.052 
 12–16 1.9 1.1 3.2 0.019 
 >16 0.9 0.4 2.0 1.000 
HLA = 17* <4 3.0 1.8 5.0 0.000 
 4–8 1.5 0.9 2.5 0.176 
 8–12 1.4 0.8 2.4 0.223 
 12–16 2.2 1.2 3.8 0.009 
 >16 0.9 0.4 2.4 1.0000 
HLA = 4 + 17* <4 2.7 1.5 5.1 0.002 
 4–8 1.5 0.8 2.8 0.239 
 8–12 1.6 0.8 3.1 0.154 
 12–16 2.2 1.1 4.2 0.033 
 >16 0.6 0.1 2.8 0.746 
HLA = 15† <4 1.7 0.2 17.3 0.509 
 4–8 0.5 0.1 4.2 1.000 
 8–12 1.0 0.1 9.2 1.000 
 12–16 2.6 0.5 14.8 0.257 
 >16 3.5 0.6 21.6 0.192 

Data are from the 1,004 patients shown in Table 2.

*

DR17 is a split of previous DR3.

†

DR15 is a split of previous DR2.

View Large

An endotype is a subtype of a condition with a pathobiological mechanism distinct from a phenotype that is an observable characteristic of a disease without any implication of a mechanism.

A recent publication mentioned several factors associated with young age of onset of type 1 diabetes as endotypes that could be evidence for heterogeneity in the disease (16). Another report described properties of pancreatic islets and low insulin secretion in patients with onset of diabetes at young ages that also support the endotype concept of type 1 diabetes (17). Our discovery of the NRP1 intron 9–encoded mechanism that lowers the percentage of insulin cells in human pancreatic islets combined with the NRP1 intron 9 genetic association with onset of type 1 diabetes at ages <2 and at 2–4 years suggests a causal factor involved in a new endotype. There is significant variation in the amount of NRP1 intron 9–encoded mRNA in isolated human pancreatic islets (Table 1) as well as in the NRP1 intron 9–encoded truncated neuropilin-1 protein isoform in islets in human pancreas (Fig. 1 and Supplementary Table 2). When a person has other risk factors for type 1 diabetes, individuals with a higher percentage of pancreatic islet cells that have the intron 9–encoded truncated neuropilin-1 isoform will have a lower mass of normal pancreatic insulin cells from birth onward. Since the net synthesis of β-cells peaks before 2 years of life and β-cell mass is set within the first 5 years of life (11), the intron 9–associated factors will, independent of other risk factors, increase susceptibility to onset of type 1 diabetes at a very young age. With the exception of the high-risk alleles of certain HLA genes, there are alleles of >30 other genes that are associated with a low effect for susceptibility to type 1 diabetes. The NRP1 intron 9-encoded mechanism and its association with the onset of type 1 diabetes at a very young age is similar to other genetic associations with low effects and further supports the endotype concept of heterogeneity in type 1 diabetes.

This article contains supplementary material online at https://doi.org/10.2337/figshare.20050979.

L.A.F. is currently affiliated with Division of Intra-abdominal Transplantation, Department of Surgery, Loyola University Medical Center and Stritch School of Medicine, Maywood, IL.

Funding. This work was supported by the Nowlin Family Trust of the InFaith Community Foundation.

The study sponsor/funder was not involved in the design of the study, the collection, analysis, or interpretation of data, or writing the manuscript and did not impose any restrictions regarding the publication of the manuscript.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.J.M. collected data. M.J.M., I.H.A., and A.S.R. designed the data analyses and analyzed data. S.W.S. and M.J.L. collected and analyzed data. J.C.E. performed the statistical analyses. P.J.C. and L.A.F. isolated and provided data on the human pancreatic islets. M.J.M. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Hasan
NM
,
Kendrick
MA
,
Druckenbrod
NR
,
Huelsmeyer
MK
,
Warner
TF
,
MacDonald
MJ
.
Genetic association of the neuropilin-1 gene with type 1 diabetes in children: Neuropilin-1 expression in pancreatic islets
.
Diabetes Res Clin Pract
2010
;
87
:
e29
–
e32
Google Scholar
Crossref
Search ADS
PubMed
 
2.
Evangelou
M
,
Smyth
DJ
,
Fortune
MD
, et al.
A method for gene-based pathway analysis using genomewide association study summary statistics reveals nine new type 1 diabetes associations
.
Genet Epidemiol
2014
;
38
:
661
–
670
Google Scholar
Crossref
Search ADS
PubMed
 
3.
Christofori
G
,
Naik
P
,
Hanahan
D
.
Vascular endothelial growth factor and its receptors, flt-1 and flk-1, are expressed in normal pancreatic islets and throughout islet cell tumorigenesis
.
Mol Endocrinol
1995
;
9
:
1760
–
1770
Google Scholar
PubMed
 
4.
Brissova
M
,
Shostak
A
,
Shiota
M
, et al.
Pancreatic islet production of vascular endothelial growth factor--a is essential for islet vascularization, revascularization, and function
.
Diabetes
2006
;
55
:
2974
–
2985
Google Scholar
Crossref
Search ADS
PubMed
 
5.
Linn
T
,
Erb
D
,
Schneider
D
, et al.
Polymers for induction of revascularization in the rat fascial flap: application of vascular endothelial growth factor and pancreatic islet cells
.
Cell Transplant
2003
;
12
:
769
–
778
Google Scholar
Crossref
Search ADS
PubMed
 
6.
Reimer
MK
,
Mokshagundam
SP
,
Wyler
K
,
Sundler
F
,
Ahrén
B
,
Stagner
JI
.
Local growth factors are beneficial for the autonomic reinnervation of transplanted islets in rats
.
Pancreas
2003
;
26
:
392
–
397
Google Scholar
Crossref
Search ADS
PubMed
 
7.
Vasir
B
,
Aiello
LP
,
Yoon
KH
,
Quickel
RR
,
Bonner-Weir
S
,
Weir
GC
.
Hypoxia induces vascular endothelial growth factor gene and protein expression in cultured rat islet cells
.
Diabetes
1998
;
47
:
1894
–
1903
Google Scholar
Crossref
Search ADS
PubMed
 
8.
Cackowski
FC
,
Xu
L
,
Hu
B
,
Cheng
SY
.
Identification of two novel alternatively spliced Neuropilin-1 isoforms
.
Genomics
2004
;
84
:
82
–
94
Google Scholar
Crossref
Search ADS
PubMed
 
9.
Mamluk
R
,
Gechtman
Z
,
Kutcher
ME
,
Gasiunas
N
,
Gallagher
J
,
Klagsbrun
M
.
Neuropilin-1 binds vascular endothelial growth factor 165, placenta growth factor-2, and heparin via its b1b2 domain
.
J Biol Chem
2002
;
277
:
24818
–
24825
Google Scholar
Crossref
Search ADS
PubMed
 
10.
Reinert
RB
,
Brissova
M
,
Shostak
A
, et al.
Vascular endothelial growth factor-a and islet vascularization are necessary in developing, but not adult, pancreatic islets
.
Diabetes
2013
;
62
:
4154
–
4164
Google Scholar
Crossref
Search ADS
PubMed
 
11.
Gregg
BE
,
Moore
PC
,
Demozay
D
, et al.
Formation of a human β-cell population within pancreatic islets is set early in life
.
J Clin Endocrinol Metab
2012
;
97
:
3197
–
3206
Google Scholar
Crossref
Search ADS
PubMed
 
12.
Gagnon
ML
,
Bielenberg
DR
,
Gechtman
Z
, et al.
Identification of a natural soluble neuropilin-1 that binds vascular endothelial growth factor: In vivo expression and antitumor activity
.
Proc Natl Acad Sci USA
2000
;
97
:
2573
–
2578
Google Scholar
Crossref
Search ADS
PubMed
 
13.
Ansari
IH
,
Longacre
MJ
,
Paulusma
CC
,
Stoker
SW
,
Kendrick
MA
,
MacDonald
MJ
.
Characterization of P4 ATPase phospholipid translocases (flippases) in human and rat pancreatic islets
.
J Biol Chem
2015
;
290
:
23110
–
23128
Google Scholar
Crossref
Search ADS
PubMed
 
14.
MacDonald
MJ
,
Hasan
NM
,
Ansari
IU
,
Longacre
MJ
,
Kendrick
MA
,
Stoker
SW
.
Discovery of a genetic metabolic cause for Mauriac syndrome in type 1 diabetes
.
Diabetes
2016
;
65
:
2051
–
2059
Google Scholar
Crossref
Search ADS
PubMed
 
15.
Olerup
O
,
Zetterquist
H
.
HLA-DR typing by PCR amplification with sequence-specific primers (PCR-SSP) in 2 hours: an alternative to serological DR typing in clinical practice including donor-recipient matching in cadaveric transplantation
.
Tissue Antigens
1992
;
39
:
225
–
235
Google Scholar
Crossref
Search ADS
PubMed
 
16.
Battaglia
M
,
Ahmed
S
,
Anderson
MS
, et al.
Introducing the endotype concept to address the challenge of disease heterogeneity in type 1 diabetes
.
Diabetes Care
2020
;
43
:
5
–
12
Google Scholar
Crossref
Search ADS
PubMed
 
17.
Leete
P
,
Oram
RA
,
McDonald
TJ
, et al.;
TIGI study team
.
Studies of insulin and proinsulin in pancreas and serum support the existence of aetiopathological endotypes of type 1 diabetes associated with age at diagnosis
.
Diabetologia
2020
;
63
:
1258
–
1267
Google Scholar
Crossref
Search ADS
PubMed
 
© 2022 by the American Diabetes Association
2022
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.