Endocrine pathology in young rabbits with cystic fibrosis

Animals

In the present study, we used wild-type (WT) rabbits and rabbits with CF, both from the New Zealand White (NZW) strain. The rabbits with CF were developed in a previous work.14 Both rabbits with CF and WT rabbits were produced by breeding heterozygous CF rabbit pairs. The rabbits were fed Laboratory Rabbit Diet (#5321, LabDiet, Richmond, Indiana, USA). All rabbits with CF were administrated polyethylene glycol (MiraLax, Bayer HealthCare, McDonough, Georgia, USA) starting at 4 weeks of age to alleviate intestinal obstruction, as previously described.14

Laboratory NZW rabbits have a lifespan of 8–10 years. They are weaned at 6 weeks of age and become sexually mature at 5–6 months of age. Based on the life cycle, rabbits are considered in ‘babyhood’ between 0 and 3 months, ‘adolescence’ between 3 and 6 months, ‘teenagers’ between 6 and 12 months, ‘young adulthood’ between 1 and 3 years, ‘middle age’ between 3 and 5 years, ‘late middle age’ between 5 and 7 years and ‘old’ if >7 years. Here, we define ‘young’ rabbits as animals between 0 and 12 months of age.

The work followed the Animal Research: Reporting of In Vivo Experiments (ARRIVE) guideline.22

H&E, Alcian blue and periodic acid-Schiff staining

The majority of the rabbit pancreases are of the mesenteric type, diffusely distributed in the mesentery of the small intestine.23 In this work, the entire mesenteric portion of the pancreas from WT rabbits and rabbits with CF were collected, and fixed with 10% neutral buffered formalin and embedded in paraffin (FFPE). The FFPE tissues were sectioned at 5 μm thickness and stained with H&E, Alcian blue (AB) or periodic acid-Schiff (PAS), as described previously.24

Immunohistochemistry staining of insulin

Immunohistochemistry (IHC) staining for insulin was performed on FFPE sections (5 μm). The slides were placed in 0.01 M citrate buffer (pH 6.0) in a pressure cooker (95°C, 2 min) for antigen retrieval, and then incubated with anti-insulin antibodies (#3014, Cell Signaling Technology, Danvers, Massachusetts, USA) overnight at 4°C. After being washed with phosphate-buffered saline, the sections were incubated with horseradish peroxidase-conjugated secondary antibody (#AS014, ABclonal, Woburn, Massachusetts, USA) at 37°C for 30 min. The sections were stained with 3,3′-diaminobenzidine (#ab64238, Abcam, Waltham, Massachusetts, USA), and slides were sealed with neutral glue. For morphometric assessment, slides were scanned using the Aperio ScanScope CS scanner (Leica Biosystems, Buffalo Grove, Illinois, USA), and images were analysed using ImageJ (V.1.54a, US National Institutes of Health, Bethesda, Maryland, USA).

Measurement of blood glucose, insulin, haemoglobin A1c and faecal elastase 1 levels

Blood glucose was measured using a Contour Next EZ Blood Glucose Monitoring System (Ascensia Diabetes Care US, Parsippany, New Jersey, USA). Insulin levels were determined using an insulin ELISA kit (#90095, Crystal Chem, Elk Grove Village, Illinois, USA).

Haemoglobin A1c (HbA1c) was measured with an A1C Now⁺ test system (#3021, PTS Diagnostics, Whitestown, Indiana, USA), which consists of one reusable analyser, a blood collector, a shaker kit with buffer solution and a test cartridge loaded with anti-HbA1c antibody. Briefly, after the blood sample was taken using the blood collector, the collector was inserted into the shaker to mix the blood with the buffer solution. The cartridge was then inserted into the analyser, and the blood sample was dispensed from the shaker into the loading area on the cartridge. The %A1c result was subsequently displayed on the analyser screen.

For the faecal elastase 1 (EL-1) assay, the daytime faecal samples from the animals were collected. Faecal EL-1 levels were measured using an ELISA kit from Immundiagnostik (#K6915, Bensheim, Germany) following manufacturer’s instructions.

Intravenous glucose tolerance test

For the intravenous glucose tolerance test (IVGTT), the rabbits were fasted for 12 hours followed by intravenous injection of glucose (0.5 g/kg body weight). Blood glucose was measured at 0, 10, 20, 30, 60, 90 and 120 min after the intravenous injection using trace amounts of blood on a test strip with a glucose monitor. At 0 and 60 min, additional blood was collected from each animal for insulin level measurement. The area under the curve (AUC) for the IVGTT data was calculated by summing the areas of successive trapezoids under the graph function using GraphPad Prism (V.10.2.2, GraphPad Software, Boston, Massachusetts, USA, www.graphpad.com).

Statistical analysis

Data were described as mean±SEM (Standard Error of the Mean) and were analysed and compared using the unpaired, two-tailed Student’s t-test with GraphPad Prism. Statistical significance was considered at p<0.05.

Rabbits with CF develop spontaneous pancreatic lesions

Previously, we reported the development of a CF rabbit line, ‘CF-9’, which carries a nine base pair deletion in the CFTR gene. The CF-9 mutation leads to the production of immature CFTR protein lacking three amino acids (P477, S478 and E479) in the nucleotide-binding domain 1.14 25 The rabbits with CF-9 exhibit many CF-like phenotypes, such as growth retardation, intestinal obstruction, dysregulated intestinal microbiome, airway abnormalities and CF-associated liver diseases.14 26 27 Here, we examined the histopathology of the pancreas in young rabbits with CF-9 (referred to as ‘rabbits with CF’ interchangeably hereafter) and WT rabbits.

H&E staining of WT rabbit pancreas revealed well-defined acinar structures with normal clusters of acinar cells, and scattered islets of Langerhans (figure 1A, A1 zoom-in box). The pancreatic ducts were lined with a single intact layer of cuboidal epithelial cells (online supplemental figure 1). In contrast, H&E staining of the pancreas of rabbits with CF showed exocrine pancreatic tissue fibrosis (figure 1B), with the loss of exocrine tissue often being replaced by connective tissues (figure 1B, B1 zoom-in box). The dilated and distended pancreatic ducts of rabbits with CF were frequently filled with eosinophilic material (figure 1B, B2 zoom-in box), and the hyperplasia of pancreatic duct epithelium was common (figure 1B, B3 zoom-in box).

Request permission

Additional micrographs of CF pancreas sections revealed further pathological features. As shown in figure 1C and the C1 zoom-in box, H&E staining indicated inflammatory cell infiltration, primarily heterophils (ie, rabbit neutrophils), in the CF pancreas. Vacuolar degeneration in the exocrine tissue was also evident (figure 1D), as well as acinar atrophy and loss of epithelial cytoplasmic zymogen granules (figure 1E). AB and PAS staining revealed metaplasia of the epithelial mucus secretory cells in the CF pancreas (figure 1F,G), but not in the WT pancreas (online supplemental figure 1).

Fatty infiltration in the exocrine pancreatic tissue was observed in one of the five (20%) rabbits with CF (online supplemental figure 2), but not in any of the WT rabbits (table 1), suggesting that this phenotype starts in rabbits with CF as early as 4 months of age. A similar observation has been reported in CF ferrets, categorised as a late-stage (phase III) lesion, in which much of the exocrine pancreas is replaced by adipocytes.28 This phenomenon has also been documented in patients with CF29 and pigs with CF.21

The pathological findings in these animals, five rabbits with CF and three WT rabbits, are summarised in table 1. Overall, most of the rabbits with CF (four out of five, 80%) exhibited pancreatic lesions, such as ductal dilation, acinar atrophy, metaplasia and vacuolar degeneration. Two rabbits with CF (two out of five, 40%) exhibited a marked loss of the exocrine pancreas, along with fibrosis and immune cell infiltration. One rabbit with CF (one out of five, 20%) showed signs of fatty infiltration. In contrast, the sections from WT rabbits (n=3) all appeared normal with minimal pancreatic lesions.

These results indicate that pancreatic lesions are common in young rabbits with CF.

Islet morphology in the pancreas of rabbits with CF

We were interested in how the insulin-secreting islets are affected in rabbits with CF. To determine if morphological changes were present in the islets of CF pancreata, we conducted immunostaining for insulin in tissue sections from rabbits with CF and compared them with those of age-matched WT rabbits. The sizes and numbers of islets were quantified. As shown in figure 2A, normal large-sized insulin-positive islet structures were observed in WT rabbit sections, covering approximately 6% of the total examined area (figure 2B). The average size of WT islets was 0.026 mm², with a largely even distribution within the 0.00 to 0.05 mm² range (figure 2C,D). In contrast, the insulin-positive islets in CF pancreatic tissue sections were much smaller, with an average size of 0.013 mm² (figure 2C). The overall insulin positive islet area was also reduced, covering only 1.5% of the total examined area, which is four times smaller than that of WT rabbits (figure 2B). Interestingly, the size distribution of CF rabbit islets was also skewed towards smaller size (figure 2D).

Request permission

Young rabbits with CF have low insulin levels

Next, we measured fasting blood glucose and insulin levels in rabbits with CF, approximately 40 days of age, and compared them with age-matched WT rabbits. The fasting blood glucose levels were similar between rabbits with CF and WT rabbits (figure 3A), consistent with our previous report.26 The HbA1c levels were also not significantly different between animals with CF and WT animals at this stage (figure 3A). However, rabbits with CF at this age had significantly lower blood insulin levels than the WT animals (figure 3A), reflecting the observed islet phenotype (figure 2).

Request permission

A subset of young rabbits with CF has compromised glucose metabolism

Approximately 20% of adolescent patients with CF develop CFRD.8 To assess whether young rabbits with CF also develop phenotypes associated with CFRD, we conducted an IVGTT in rabbits with CF (n=10, average age on the test day: 48.5±3.3 days) and WT (n=10, average age on the test day: 52.8±7.6 days) rabbits. Overall, the blood glucose levels of rabbits with CF had a slower return-to-normal rate after glucose administration compared with WT rabbits (figure 3B), suggesting that an indeterminate glucose tolerance (INDET) phenotype, a pre-CFRD stage characterised by elevated AUC despite a return-to-normal glucose level at the end of the test, exists in some rabbits with CF. Consistently, the AUC of the IVGTT response in rabbits with CF was larger than that in the WT rabbits (figure 3B).

We also compared the insulin levels, and relative fold-changes of insulin levels in these animals at 0 min (ie, prior to glucose administration) and 60 min postglucose administration (online supplemental figure 3A). The data confirmed that rabbits with CF in general have lower insulin levels compared with WT rabbits. Furthermore, the data suggest that rabbits with CF secrete less insulin in response to glucose stimulation, as evidenced by the smaller fold-change in insulin levels (calculated as the ratio of insulin levels at 60 min to those at 0 min) in the animals with CF compared with WT animals (online supplemental figure 3B). This indicates that rabbits with CF have impaired glucose-stimulated insulin secretion, a known condition in patients with CF.30

We then stratified the rabbits with CF into two groups based on their IVGTT responses. The first group consisted of six animals (age on the test day: 47.5±3.0 days), whose IVGTT curves were indistinguishable from those of WT animals (figure 3C). The second group consisted of four animals (age on the test day: 50.0±3.0 days), whose IVGTT curves diverged from those of the WT animals starting from the first postglucose measurement time point (ie, 10 min). We referred to these animals as ‘INDET-like’. The AUC of the IVGTT curves from these four INDET-like animals was significantly higher than those of the other animals with CF (figure 3C), and not surprisingly, also significantly higher than those of the worst four (WT-worst-4, defined as the four animals with the highest AUC values) of the WT, and the best six (WT-best-6, defined as the six animals with the lowest AUC values) of the WT (online supplemental figure 3C).

It has been reported that in the age group of 30–39 years, women with CFRD outnumber men.31 Therefore, we investigated whether there is a sex effect on the glucose metabolism in rabbits with CF. Interestingly, all INDET-like rabbits with CF were female (n=4). Of the six female rabbits with CF (average age on the test day: 50.0±1.2 days) analysed, four exhibited INDET-like phenotypes (4/6, 67%), while only two (2/6, 33.3%) had normal IVGTT responses. In contrast, all male rabbits with CF (n=4, average age on the test day: 46.3±1.1 days) had normal IVGTT responses. Consistently, the IVGTT and the AUC of female and male rabbits with CF were significantly different (figure 3D), whereas no significant difference was noted between WT female and male animals (online supplemental figure 3D,E).

These data indicate that a subset of young rabbits with CF develop INDET phenotypes, and that female rabbits with CF are more prone to developing INDET than male ones at this age. The results also demonstrate that rabbits with CF reflect the heterogeneity and complexity of CF-related pancreatic endocrine dysfunction.

EL-1 levels in daytime faeces of rabbits with CF

In the clinic, PI is determined by an ELISA assay to measure the faecal EL-1 levels.32 A near-zero faecal EL-1 value indicates PI, and a high EL-1 value suggests otherwise. Here, we employed the ELISA assay to assess the EL-1 levels in the daytime faeces of WT rabbits and rabbits with CF. Interestingly, while the daytime faecal EL-1 levels of rabbits with CF (8 of 8, 100%) were all below the detection limit, suggesting a very low EL-1 activity, the readings from four of the eight (50%) WT faecal samples were also very low (online supplemental figure 4). This may be attributed to the fact that rabbits produce two types of faecal material: a hard, dry ‘daytime’ faecal pellet and a soft protein-rich ‘night’ faeces known as cecotropes, which are consumed directly by the rabbit as they are excreted at night.33 In the present work, we were unable to collect the cecotropes, and all faeces assayed were daytime samples.

As such, the EL-1 results derived from daytime faeces did not provide conclusive support to the PI phenotype in rabbits with CF; measuring EL-1 levels in the nighttime cecotropes of rabbits with CF will be necessary in a future study.