BPIFB4 protein constructs
Hek-293 cells were transfected with the LAV-BPIFB4 vector cloned in fusion with His-Tag. The recombinant protein was purified using affinity Nuvia IMAC Resin (Bio-Rad). As a control, we used the eluate of cells transfected with the null vector.35
In vivo studies
Two controlled, randomized studies followed the EU Directive 2010/63/EU and the principles stated in the Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, 1996). The procedures detailed below were prepared with support from the Experimental Design Assistant, a free resource from the National Centre for Replacement, Refinement, and Reduction (3Rs) of Animals in Research 166 (https://eda.nc3rs.org.uk/), were covered by an ethical license approved by the British Home Office and the Universities of Bristol and Cardiff (# PP1377882, valid until October 25, 2025), and they were deposited, before starting the experimentation, on The Animal Study Registry (Title: Transferring healthy longevity gene to improve age-related heart dysfunction, DOI: 10.17590/asr.0000316, Date of registration:2023-04-26, www.animalstudyregistry.org). The principal investigator (PM) will provide access to the deposited protocol and experimental data upon request.
Maintenance and randomization. The mice were housed in groups of 1–6 individuals (as required by the experimental procedure) in an enriched environment within a biosecure unit under a 12 h light/dark cycle, fed with EURodent Diet (5LF5, LabDiet, Durham, UK) and given drinking water ad libitum. The GraphPad software (https://www.graphpad.com/quickcalcs/randomize) was used to randomly assign the subjects to the control and treatment groups (fixed ratio = 1:1).
Methods to minimize bias. The LAV-BPIFB4 protein and vehicle were administered by an investigator blinded to the intervention status (VVA). Similarly, the echocardiography and histological analyses were performed by investigators (VVA, YQ, and SS) who were blinded to the randomization. The animals were exposed to the same enriched environment and received the same care and observation to avoid performance bias. An intention-to-treat approach was applied to maintain the integrity of randomization and strengthen the trial's internal validity. In line with this, participants in the two arms of the study (vehicle or LAV-BPIFB4 protein) were considered in the analysis of the echocardiography data regardless of whether they did not complete the follow-up due to premature termination.
Treatment. Following baseline echocardiography (Vevo 3100, VisualSonics), 18-month-old C57BL/6J mice (Charles River, UK) were randomized to receive a recombinant purified LAV-BPIFB4 protein (100 µL of 3 µg/ animal) or an equivalent volume of vehicle by gavage every 3 days under isoflurane anesthesia (2–3%). Our previous study using LAV-BPIFB4 gene therapy did not reveal any difference in the treatment effect between older male and female mice. Therefore, the present LAV-BPIFB4 protein therapy was conducted in male mice only. On the other hand, whether sex influences LAV-BPIFB4 therapy under diabetic conditions has not been previously tested.23 Therefore, the same protocol for LAV-BPIFB4 protein or vehicle administration was used for 9-week-old, male and female C57BLKS/J-Leprdb/Leprdb/Dock7+ [db/db] mice. Animals were terminated at the 30-day follow-up after a final echocardiography assessment under isoflurane anesthesia by exsanguination, followed by removal of the heart for histology. In a pilot study on elderly mice, we found that a 2-week treatment was sufficient to appreciate the benefit of LAV-BPIFB4 therapy on echocardiography and histology endpoints. Nonetheless, we decided to extend the follow-up to 30 days to maximize the information on efficacy and cross-compare our findinges with those of our LAV-BPIFB4 gene therapy studies, which had the same duration.22,23 Previous reports have shown that repeated four-week gavages are well tolerated and do not cause metabolic effects in leptin-deficient Ob/ob mice with type 2 diabetes.28,29
Endpoints. Primary: indices of cardiac function; secondary: microvascular density, fibrosis, and senescence.
Determination of glucose levels
Blood glucose was determined at the end of the study via a blood glucose monitor. Since this machine's range is up to 34 mmol/l, and blood glucose levels above this range are high (Hi), at termination, urine glucose was quantified via a glucose colorimetric assay kit according to the manufacturer's instructions (Abcam #ab65333).
Tissue collection and histological analysis
The heart was stopped in diastole with 30 mM (w/v) KCl, flushed with 1% EDTA in normal phosphate-buffered saline (PBS) and cut into pieces; the top section was flash-frozen for future protein and molecular biology studies, and the lower section was drop-fixed in 4% PFA. All the PFA samples were incubated at + 4°C for 18–24 h, after which the PFA was replaced with PBS. The PFA-fixed cardiac tissues were cryoprotected with 30% sucrose for 24–48 hours before being embedded in OCT compound. Histochemical and immunohistochemical studies were performed on six-µm-thick sections cut with a Thermo Fisher Scientific CryoStar NK50 cryostat unless otherwise stated.
Histological analyses of mouse hearts. The specific antibodies and procedures used are listed in Supplementary Table 1. All the immunohistochemical procedures included tissue sections without primary antibodies as a technical internal control and suitable irrelevant IgG as a negative control. Images were acquired via a transmitted light microscope (Olympus BS40), a confocal microscope (Leica TCS-SP8), or an epifluorescence microscope (Zeiss AxioObserver Z1).
Vascular density was measured by counting capillaries and arterioles (Zeiss Observer. Z1 microscope, 20x objective). The final data are expressed as the number of capillaries and arterioles per mm2. Analysis of PDGFRβ + pericytes (PCs) associated with the coronary capillaries and arterioles was carried out on sections stained with anti-PDGFRβ antibodies to identify PCs overnight at + 4°C (1:50, R&D Systems, #AF1042), with alpha-smooth muscle actin (α-SMA, 1:400, Sigma, #C61198), arteries identified with isolectin-B4 (1:200, Life Technologies, #121414), endothelial cells identified with α-sarcomeric actin (α-SA, 1:200, IgM isotype, Sigma, #A2172) overnight at + 4°C. For analysis of PC density, images with capillaries in cross-sections were obtained via a 20x objective. The density of PCs is expressed as the number of IB4- and α-SMA PDGFRβ + perivascular cells/mm2 of the α-SA + myocardial area considered separately pericytes around capillaries and arterioles. The cryosections were stained for collagen via the Azan Mallory method (Heidenhain’s adaptation of Mallory’s trichrome stain). The degree of fibrosis was assessed in the perivascular area and quantified as the ratio of collagen to vessel area. Interstitial fibrosis was also measured in pixels and expressed as a percentage of the tissue area.
A senescence β-galactosidase staining kit (Cell Signaling Technology #9860) was used to identify senescent cells in the frozen cardiac sections (Olympus microscope, ×400 magnification). In addition, senescent cardiomyocytes and interstitial cells were recognized via anti-mouse P16ink4A (1:50, Santa Cruz, #sc-1661) and expressed as the number of positive nuclei per mm2 of tissue. Cardiac muscles were stained with α-SA for 2 hours at RT, and Alexa Fluor 647-conjugated anti-mouse IgM (1:200, for 1 hour, at + 20°C; Life Technologies, UK) was used as a secondary antibody. Slides were stained with 1:1000 (v/v) DAPI solution in 1x PBS and mounted with Fluoromount G for imaging. Representative and quantitative images were taken using a Zeiss Observer. A Z1 microscope was set up on a fluorescent field path with a 20x objective. P16ink4A-positive cells are expressed as the number of positive nuclei per mm2 of tissue. According to the manufacturer's instructions, cardiac samples were assessed for apoptosis via a TUNEL assay (ApopTag® Red In Situ Apoptosis Detection Kit, Millipore #S7165).
For BPIFB4 staining, antigen retrieval was performed on cryosections using citrate buffer (pH = 6; 1x; Sigma‒Aldrich) for 15 min at + 98°C to unmask the epitopes within the tissue. A blocking solution containing 5% v/v goat serum was used to stop nonspecific binding. Antibody staining was performed using an anti-α-SA antibody (1:200) for two hours at RT to recognize cardiac muscle. Alexa Fluor 647-conjugated anti-mouse IgM (1:200) was used as a secondary antibody. A rabbit anti-BPIFB4 antibody (GeneTex, #GTX51455) was used as the primary antibody at a ratio of 1:100, and Alexa Fluor 568-conjugated anti-rabbit antibody was used at a ratio of 1:200 for BPIFB4 visualization. Nuclei were labeled with DAPI (1:1000), and slides were covered with aqueous Fluoromount G and coverslips for fluorescence imaging. Images were taken via a Leica TCS-SP8 confocal laser scanning microscope attached to a Leica DM I8 inverted epifluorescence microscope and processed via LASX software. Lightning adaptive image restoration was used to optimize the resolution. ImageJ Fiji was used to merge the fluorescent-colored channels and to add scale bars. The staining quantification was performed on the whole heart area and its subregions (inner, middle, and outer cardiac layers), and the results are illustrated separately. All the morphometric analyses were carried out with ImageJ software.
Statistical analyses.
The data are presented as individual values and means ± standard errors of the means (SEMs) or standard deviations. The D’Agostino‒Pearson and Kolmogorov‒Smirnov normality tests were used to check for a normal distribution when applicable. Echocardiography parameters (baseline and final assessed in the same animal) were evaluated via a mixed-effects model, 2-two-way ANOVA, which was used to examine the influence of two different categorical independent variables (treatment and sex) on one continuous dependent variable and then to determine if they were interacting. Post hoc analyses included Tukey’s comparison test. The unpaired t-test was used for all other analyses to compare two groups, except for data that were not normally distributed, for which the Mann‒Whitney U test was used. Significance was assumed when the P value was 0.05 or less. Analyses were performed via GraphPad Prism 10.0 (San Diego, CA, USA).