Validation and Preclinical Correlation of a New Sandwich ELISA for Measuring Murine Hepcidin.

Abstract 2100

Although enzyme-linked immunosorbent assays (ELISA) exist for measuring human hepcidin, no commercially available antibodies to mouse hepcidin are validated for this purpose. Thus, we developed an ELISA to measure mouse hepcidin using low plasma volumes to allow for repeat testing of individual mice over time. To this end, human hepcidin knock-in mice (Hep2 mice) were used to generate a panel of anti mouse hepcidin antibodies. A sandwich ELISA using these antibodies was validated. Pooled plasma from iron-deficient (combining diet and phlebotomy), iron-replete, and inflamed (injecting sub-clinical amounts of lipopolysaccharide) FVB mice was used to prepare frozen aliquots of low, medium, and high reagent control plasma. These samples were tested on 5 different days (n=8 wells/sample/day). Mean hepcidin levels for the high, medium, and low reagent were 791, 504, and 117 ng/mL, respectively. The intra-assay coefficient of variation (CV) was 10, 4.5, and 15.6% for the high, medium, and low samples, respectively. Inter-assay CV was 4.7, 4.0, and 11.1% for the high, medium, and low samples, respectively. Analytical sensitivity was 0.7 ng/mL and the assay was linear (Pearson r=0.99) to 500 ng/mL at a 1:100 dilution of sample to blocking buffer (i.e. only 1 uL of sample needed per test). With 20 plasma samples spanning the range of expected hepcidin levels, there was high correlation between hepcidin determined by ELISA and the sum of hepcidin-1 and hepcidin-2 by time-of-flight mass spectrometry (Pearson r = 0.99; P<0.0001). Stability was analyzed using samples from repeated freeze-thaw cycles at −80°C or from incubation at 4°C. Each freeze-thaw cycle induced up to a 15% loss in hepcidin. Overnight refrigeration led to a 31%, 21%, and 4% decrease in hepcidin for the high, medium, and low samples, respectively. Severe hemolysis (i.e. 8 g/dL hemoglobin) caused interference, decreasing apparent hepcidin levels up to 52%; mild hemolysis (0.1 g/dL hemoglobin) decreased hepcidin levels up to 15%. To provide clinical correlation in mouse models, we used cohorts (n=5/group) of C57BL/6 females with differing iron status:

1. Iron deficient: weanling mice on a low iron diet (0–5 ppm FeSO₄; AIN diet) for 8 weeks

2. Iron-replete: weanling mice on a normal iron diet (45 ppm FeSO₄; AIN diet) for 8 weeks

3. Iron-overload: mice received weekly iron-dextran injections (5 mg for 7 weeks)

4. Transfusional iron-overload: mice received weekly syngeneic RBC transfusions (1 unit for 7 weeks)

Hepcidin mirrored ferritin levels and total liver iron measured at necropsy (see Table). To test whether this ELISA can track hepcidin during infection, and whether iron status affects this response, iron-deficient or -replete mice (n=10/group) were inoculated with 1000 colony forming units of Salmonella typhimurium and blood samples obtained pre-infection and 1–7 days post-infection to measure hepcidin and interleukin (IL)-6. Pre-infection, iron-deficient mice had significantly lower hepcidin levels than iron-replete mice (<0.7 vs. 136 ± 56 ng/mL, mean ± sd). On Day 7 post-infection, IL-6 was much higher in surviving iron-deficient mice (up to >2000 pg/mL) than in iron-replete mice (207 ± 70 pg/mL). Despite dramatic increases in IL-6, hepcidin increased less in iron-deficient mice (260 ± 132 ng/mL) than in iron-replete mice (636 ± 273 ng/mL). Thus, despite a stronger inflammatory stimulus, hepcidin levels are muted in iron-deficient animals. In conclusion, this validated ELISA is a new tool that can be used to study iron regulation in mice.