Neurofilament light: a possible prognostic biomarker for treatment of vascular contributions to cognitive impairment and dementia

Abstract

Background: Decreased cerebral blood flow and systemic inflammation during heart failure (HF) increase the risk for vascular contributions to cognitive impairment and dementia (VCID) and Alzheimer disease-related dementias (ADRD). We previously demonstrated that PNA5, a novel glycosylated angiotensin 1-7 (Ang-(1-7)) Mas receptor (MasR) agonist peptide, is an effective therapy to rescue cognitive impairment in our preclinical model of VCID. Neurofilament light (NfL) protein concentration is correlated with cognitive impairment and elevated in neurodegenerative diseases, hypoxic brain injury, and cardiac disease. The goal of the present study was to determine (1) if treatment with Ang-(1-7)/MasR agonists can rescue cognitive impairment and decrease VCID-induced increases in NfL levels as compared to HF-saline treated mice and, (2) if NfL levels correlate with measures of cognitive function and brain cytokines in our VCID model.

Methods: VCID was induced in C57BL/6 male mice via myocardial infarction (MI). At 5 weeks post-MI, mice were treated with daily subcutaneous injections for 24 days, 5 weeks after MI, with PNA5 or angiotensin 1-7 (500 microg/kg/day or 50 microg/kg/day) or saline (n = 15/group). Following the 24-day treatment protocol, cognitive function was assessed using the Novel Object Recognition (NOR) test. Cardiac function was measured by echocardiography and plasma concentrations of NfL were quantified using a Quanterix Simoa assay. Brain and circulating cytokine levels were determined with a MILLIPLEX MAP Mouse High Sensitivity Multiplex Immunoassay. Treatment groups were compared via ANOVA, significance was set at p < 0.05.

Results: Treatment with Ang-(1-7)/MasR agonists reversed VCID-induced cognitive impairment and significantly decreased NfL levels in our mouse model of VCID as compared to HF-saline treated mice. Further, NfL levels were significantly negatively correlated with cognitive scores and the concentrations of multiple pleiotropic cytokines in the brain.

Conclusions: These data show that treatment with Ang-(1-7)/MasR agonists rescues cognitive impairment and decreases plasma NfL relative to HF-saline-treated animals in our VCID mouse model. Further, levels of NfL are significantly negatively correlated with cognitive function and with several brain cytokine concentrations. Based on these preclinical findings, we propose that circulating NfL might be a candidate for a prognostic biomarker for VCID and may also serve as a pharmacodynamic/response biomarker for therapeutic target engagement.

Keywords: Angiotensin-(1–7); Biomarker; Inflammation; Neurofilament light (NfL); PNA5; Vascular contributions to cognitive impairment and dementia.

Conflict of interest statement

COI, Dr. Meredith Hay is the founder of ProNeurogen, Inc. that holds exclusive rights to the clinical development of the Ang-(1–7)/MasR agonists described herein.

© 2021. The Author(s).

Figures

Fig. 1

Experiment timeline. Three-month-old male mice were acclimated for one week before MI or Control surgeries. 5 weeks following recovery, mice were treated with Saline, Ang-(1–7) or PNA5 at 2 concentrations (50 and 500 ug/kg) for 24 days. During the last 3 days of treatment mice underwent Novel Object Recognition testing. Following Novel Object Recognition testing, mice underwent echocardiographs before killing. At killing, the plasma, brain, and heart were collected. MI myocardial infarct, S.C. subcutaneous injection

Fig. 2

Echocardiograms and histology to confirm myocardial infarctions in the heart failure model. A Simpson Ejection Fraction was calculated using V-mode left ventricle calculations treatment groups: heart failure (HF)-A50 (Ang-(1–7)50 ug/kg), HF-P50 (PNA5 50ug/kg), HF-A500 (Ang-(1–7) 500 ug/kg), HF-P500 (PNA5 500 ug/kg), HF-Saline, Control-Saline. Significant difference between Control-Saline and the following treatment groups are as listed: HF-Saline, HF-A50 (Ang-(1–7)50 ug/kg), HF-P50 (PNA5 50 ug/kg), HF-A500 (Ang-(1–7)500 ug/kg), HF-P500 (PNA5 500 ug/kg), P < 0.05 ANOVA, Dunnett’s post. B Simpson End Systolic Volume was calculated using V-mode left ventricle calculation. Significant difference between Control-Saline and the following treatment groups are as listed: HF-Saline p = 0.001, HF-P50 (PNA5 50 ug/kg) p = 0.0001, and HF-A500 (Ang-(1–7)500 ug/kg) p = 0.015, ANOVA, Dunnett’s post. C Examples of hematoxylin and eosin staining of paraffin embedded formalin fixed hearts demonstrates differences in wall morphology between Control animals and HF-saline, PNA5, and Ang-(1–7) treatment mice. Left ventricle wall is indicated with black arrow. Images were taken at 5 × magnification

Fig. 3

Ang-(1–7)/MasR agonists rescue cognitive impairment in a VCID mice. A Discrimination ratio = (Time spent at new object- Time spent at familiar object)/ total exploration time. Heart failure (HF) mice were treated with saline PNA5 or Ang-(1–7) at two different concentrations of either 50ug/kg, (P50, A50) or 500ug/kg (P500, A500). ****p < 0.001 Control-Saline vs HF-Saline. *p < 0.05 HF-Saline vs HF-P50, HF-P500, HF-A500. Data are represented with mean and ± SE. Difference in groups were tested by ANOVA, Dunnett’s post, significance p < 0.05. B No significant differences are noted in total exploration time in the familiarization test for all treatment groups. Data are represented with mean and ± SE. Significancy were tested by ANOVA

Fig. 4

Treatment with Ang-(1–7)/Mas R agonists decreases plasma NfL levels. A Data are represented with mean and ± SE. Compares NfL levels between control saline and HF-saline mice, p = 0.08, Welch’s t-test. This difference did not reach significance. B PNA5 at 50ug/kg and Ang-(1–7) at 50 ug/kg, and 500 ug/kg significantly reduced NfL plasma concentration in VCID mice compared to HF-saline treated mice, Kruskal–Wallis ANOVA, Dunn’s post, p < 0.05. C NfL plasma levels significantly negatively correlated with Discrimination ratios. Correlation R values were produced using Pearson’s correlation analysis and the fit line was determined using simple linear regression

Fig. 5

Increased TNFα in VCID models are inhibited by PNA5 and Ang-(1–7). Plasma TNFα of Control-Saline, HF-Saline, HF-A50 (Ang-(1–7)50 ug/kg), HF-P50 (PNA5 50ug/kg), HF-A500 (Ang-(1–7)500ug/kg), HF-P500 (PNA5 500ug/kg), were measured via MILLIPLEX MAP Mouse High Sensitivity Multiplex Immunoassay. Significance were observed between HF-Saline vs Control-Saline treated mice (3.50± SE 0.15, n = 6 vs mean 3.12, ± SE 0.01, n = 6, respectively, p = 0.003;), PNA5, and Ang-(1–7) (HF-P50 mean 3.12 pg/mL± SE 0.00, n = 5 p = 0.004; HF-A50 mean 3.12 pg/mL± SE 0.00, n = 6, p = 0.003; HF-P500 mean 3.12 pg/mL± SE 0.00, n = 6, p = 0.003; ANOVA, Dunnett’s post; values that were at or below assay threshold of 3.12 pg/mL were represented as 3.12 pg/mL)

Fig. 6

Effect of treatment with Ang-(1–7)/MasR agonists on brain cytokine levels. Whole brain lysate cytokine levels were measured using MILLIPLEX MAP Mouse High Sensitivity Multiplex Immunoassay. A–F HF-Saline treated VCID mice had significantly lower IL-1α, IL-2, IL-17, and IP-10 levels in comparison to Control-Saline treated mice. PNA5 and Ang-(1–7) significantly increased cytokines levels as compared to HF-saline treated animals. Data are represented with mean and ± SE. Differences in groups were tested by ANOVA, Dunnett’s post hoc, significance set a p < 0.05

Fig. 7

Relationship between brain cytokine levels and NfL levels. A–E The relationship between brain cytokines IL-2, IL-5, IL-13, IL-17, IP-10, and plasma NfL levels were analyzed using Pearson correlation. F Brain cytokine composite score of IL-2, IL-13, IL-17 were significantly negatively correlated with NfL plasma concentrations

Fig. 8

Pearson’s correlation matrix of the relationship between plasma NfL and brain cytokine levels. The relationship between the brain cytokines IL-2, IL-5, IL-13, IL-17, IP-10, the composite cytokine z-score and plasma NfL levels were analyzed using Pearson correlation. Blue demonstrates positive correlation and red represents negative correlation. Values closer to 1 or -1 indicate more closely correlated variables

Fig. 9

Longitudinal comparison of plasma cytokines in VCID mouse model. A–C Plasma inflammatory cytokine levels from mice with 3 weeks of HF, and 8 weeks of HF, in both Control-Saline, and HF-Saline treated mice were measured via MILLIPLEX MAP Mouse High Sensitivity Multiplex Immunoassay. At 3 weeks of HF, HF-Saline mice had significantly higher IL-1α, MIP-1α, and MIP-2 concentrations than age-matched Control-Saline treated mice. A significant decrease in cytokine concentrations was observed in HF-Saline-treated mice from 3 to 8 weeks of HF. Data are represented with mean and ± SE. Difference in groups were tested by ANOVA