Insulin-Like Growth Factor 1 in the Preterm Rabbit Pup: Characterization of Cerebrovascular Maturation following Administration of Recombinant Human Insulin-Like Growth Factor 1/Insulin-Like Growth Factor 1-Binding Protein 3

Following preterm birth, serum levels of insulin-like growth factor 1 (IGF-1) decrease compared to corresponding in utero levels. A recent clinical trial indicated that supplementation with recombinant human (rh) IGF-1/rhIGF-binding protein 3 (rhIGF-1/rhIGFBP-3) prevents severe intraventricular hemorrhage (IVH) in extremely preterm infants. In a preterm rabbit pup model, we characterized endogenous serum and hepatic IGF-1, along with brain distribution of IGF-1 and IGF-1 receptor (IGF1R). We then evaluated the effects of rhIGF-1/rhIGFBP-3 on gene expression of regulators of cerebrovascular maturation and structure. Similar to preterm infants, serum IGF-1 concentrations decreased rapidly after preterm birth in the rabbit pup. Administration of rhIGF-1/rhIGFBP-3 restored in utero serum levels but was rapidly eliminated. Immunolabeled IGF1R was widely distributed in multiple brain regions, displaying an abundant density in the choroid plexus and sub-ependymal germinal zones. Increased IGF-1 immunoreactivity, distributed as IGF1R, was detected 4 h after rhIGF-1/rhIGFBP-3 administration. The rhIGF-1/rhIGFBP-3 treatment led to upregulation of choroid plexus genes involved in vascular maturation and structure, with corresponding protein translation for most of these genes. The preterm rabbit pup model is well suited for evaluation of IGF-1-based prevention of IVH. Administration of rhIGF-1/rhIGFBP-3 affects cerebrovascular maturation, suggesting a role for it in preventing preterm IVH.

approx. 3 hours of age, the pups received a 0.1 ml s.c. dose of 8.0 mg/kg rhIGF-1/rhIGFBP-3 or vehicle (9 mg/ml NaCl, B Braun) that was repeated every 12 hours up to approx. 72 hours post administration (one (1) to seven (7) administrations of rhIGF-1/rhIGFBP-3 or vehicle in total). At approx. 0 (0), 4 (7), 12 (15), 24 (27), 48 (51), and 72 (75) hours post rhIGF-1/rhIGFBP-3 or vehicle administration (corresponding postnatal age within brackets), the pups were terminated and samples collected for choroid plexus gene expression analysis according to the following procedure: At the respective timepoints, the pups were sedated as described above, followed by opening of the chest and apical cardiac puncture perfusion with 50 ml of 0.1 M phosphate buffered saline (PBS, pH 7.4) for 30 seconds (until visible bleaching of the tissue was observed). After completion of perfusion, the pups were terminated by cervical dislocation and the brains dissected out from the skull and placed in an ice-cold petri dish containing 6 ml Krebs-Ringer HEPES solution (Alfa Aesar, Kandel, Germany). By sharp dissection, the hemispheres were divided and the choroid plexus removed from the ventricular cavity by blunt dissection with a pincette. The choroid plexus tissue biopsy, as well as a liver and ear biopsy samples collected at termination, were snap frozen on dry ice and processed as described below. From all animals, a blood sample was collected at the respective termination time-points and processed as described below. At approx. 4 (7), 24 (27), 48 (51), and 72 (75) hours post rhIGF-1/rhIGFBP-3 or vehicle administration (corresponding postnatal age within brackets), the pups were terminated and samples collected for histochemistry and immunohistochemistry analysis according to the following procedure: Following sedation as described above, the pups were perfused with 0.9% saline, followed by transcardial perfusion with freshly prepared 4% PFA, and the brains were dissected out from the skull, immersed in 4% PFA, and processed as described below.

Tissue collection and processing for histology and immunolabeling
Following fixation, the brain tissue was dehydrated in a graded ethanol series (70-99.99%), followed by xylene (100%), and finally immersed in paraffin and embedded in paraffin blocks. Coronal and sagittal brain sections (4 or 6 µm) were prepared on a rotating microtome (Microm HM 360, Microm International GmbH, Walldorf, Germany), and sections were collected on microscope slides (SuperFrost Plus, Thermo Scientific/Gerhard Menzel B.V. & Co., Braunschweig, Germany).

Immunohistochemistry
The specific labeling of the primary antibodies against IGF1R and IGF-1 was confirmed by the distribution of labeling in rabbit brain tissue, the distribution of labeling in rabbit (not shown) and human placenta tissue (i.e. positive control tissue), and the lack of labeling when excluding the primary antibody from the protocol (i.e. negative control) ( Supplementary Fig. 2). The primary antibody used against IGF1R, a polyclonal goat anti-IGF1R (AF-305-NA #639h, R&D System, McKinley Place, MN, USA), provided strong labeling with low background (using both chromogen immunohistochemistry and immunofluorescence). The primary antibody used against IGF-1, a monoclonal mouse anti-IGF-1 (AM33345PU-S, Origene, Herford, Germany), provided a strong but less distinct labeling. For chromogen visualization of IGF-1 immunoreactive sites, nickel ion intensification of the immunoreaction was required.
The chromogen labeled coronal and sagittal sections were analyzed by visual inspection of microscope slides in a bright-field microscope (Olympus IX73, Shinjuku, Tokyo, Japan). Whole sections were scanned (Hamamatsu, NanozoomerS60, Hamamatsu, Japan), and digital image processing was performed to visualize the IGF1R neuroanatomical distribution and differences in receptor densities in the brain. An ImageJ macro was developed, and the threshold for detection of IGF1R immunoreactivity was set from the control slides (no primary antibody incubation). For quantitative analysis and comparison between animal groups, the IGF1R density in digital images of the IGF1R immunohistochemistry labeling in whole coronal sections (4-7 sections per animal) was analyzed.
Artifacts (reaction precipitates and damaged tissue) were extracted from the images. Using the set threshold value, IGF1R densities per tissue area were measured, and the percentage of labeling was compared between the groups (not shown). Color coded images were made for illustrations of the differentiated receptor densities in the brain, represented as high (red), medium (green), low (blue), and no staining (black) (see Fig. 2A-B, Supplementary Fig. 2).

Immunofluorescence
Immunofluorescence labeling was used to elucidate the detailed distribution of IGF1R and for evaluation of the relation between IGF1R and IGF-1, endogenous and/or from administered rhIGF-1/rhIGFBP-3, as well as IGF-1 and the endothelial cell marker CD31. Single (IGF1R) or double labeling (IGF1R vs. IGF-1 or IGF-1 vs. CD31) was performed by using the IGF1R and IGF-1 antibodies as described above and a goat CD31 polyclonal IgG antibody (2 µg/ml diluted in PBS-TX-BSA, sc-1505, Santa Cruz Biotech, Dallas, TX, USA).
Immunofluorescence labeling was also used for detection and visualization of proteins selected from the gene expression analysis following exposure to rhIGF-1/rhIGFBP-3. The following primary antibodies were used: angiopoietin-1 (ANGPT1, goat polyclonal IgG, 2 µg/ml diluted in PBS-TX-BSA, Double labeling of IGF1R and IGF-1 and of IGF-1 and CD31 was performed as antibody cocktails, i.e. incubation with a mixture of primary antibodies and secondary antibodies, respectively. Sections were then rinsed in PBS-TX and incubated in 4′,6-diamidino-2-phenylindole (DAPI) for 15 minutes at room temperature, followed by rinsing in PBS and mounting in anti-fade solution (Abcam). In every run, primary antibody specificity was tested by omitting the primary antibody incubation from the protocol on adjacent sections to those incubated with primary antibodies (i.e. negative control, not shown).
These control sections were also used to set the background threshold level for fluorescence detection, i.e. only fluorescence signal levels above the threshold were recorded. Confocal laser scanning analysis and imaging of the fluorescence labeling were performed using a Zeiss LSM 800 microscope (Zeiss, Germany) with x20 dry or x40 oil immersion lenses. Sequential scanning was performed with an optimized pinhole (airy unit) for each wavelength/fluorophore. For the double labeling analysis, the pinholes were optimized for all channels to ensure detection and visualization of the individual labeling at the same focal plane. For quantitation of ANGPT1, FN1, PROCOL, VCAN, and THBS1, confocal images were grabbed of the region of interest, i.e. choroid plexus (4-8 images per animal) and exported to TIFF format. The area of interest was specifically selected and annotated manually. An ImageJ macro was developed to extract the immunofluorescence labeled targets, providing the total analyzed area and the DAPI labeled cell nuclei area. The immunofluorescence labeled area was measured and calculated per cell nuclei area (percentage), which was used for comparison between rhIGF-1/rhIGFBP-3 and vehicle treated animals. Amplification was performed as described by the manufacturer (Qiagen, RT2 SYBR Green qPCR CFX Connect (Bio-Rad). Data were analyzed using Bio-Rad CFX Maestro 1.1 (Bio-Rad).

Sex determination
Rabbit sex was determined by confirmation of the presence of the sex determining region Y gene (SRY, gene ID: 100328958) sequence in the rabbit genome using PCR and visualization by gel electrophoresis. DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen) according to the manufacturer's instructions. One (1) µl of DNA was used in respective PCR reaction (30 cycles at 57°C) with the following primers: Sense: TGCAATACAGGAGGAACACG, Antisense: AGCAAACTGTCGCTCTTCTG. The presence of a band at approx. 299 bp was determined as male, and correspondingly no visual band was determined as female. Digital image processing of IGF1R chromogen labeled (brown, left) whole sections was performed to elucidate the IGF1R neuroanatomical distribution and differences in receptor densities in the brain, and color coded images (right) were made for illustrations of the differentiated receptor densities in the brain, represented as high (red), medium (green), low (blue), and no staining (black).

Supplementary Tables
Supplementary Table 1 Sample was obtained at termination, i.e. one serum IGF-1 value at one timepoint was obtained per animal. Timepoint is presented as time after rhIGF-1/rhIGFBP-3 administration (corresponding age of pup at sampling/termination). 3 Data is presented as median (range, n). Sample was obtained at termination, i.e. one liver mRNA IGF-1 value at one timepoint was obtained per animal. Timepoint is presented as time after rhIGF-1/rhIGFBP-3 administration (corresponding age of pup at sampling/termination). 3 Data is presented as median fold change (range, n) vs. expression levels at T=0h (E29 + 0h). Data was normalized to ACTB.