Association between Mid-Regional Proadrenomedullin Levels and Progression of Deep White Matter Lesions in the Brain Accompanying Cognitive Decline

INTRODUCTION

In a super-aged society, the possibility of early discovery and prevention of vascular dementia using novel surrogate markers has been attracting increasing attention. Cognitive decline accompanied by ischemic cerebral deep white matter lesions (DWMLs) is a typical finding in subcortical vascular dementia [1], and its fundamental pathology is arteriosclerosis of the medullary artery in the white matter and the subsequent impairment of cognitive function due to diffuse DWMLs [2]. In addition to classic risk factors such as hypertension and arteriosclerosis, a number of other factors are involved in the development of DWMLs [3], including endothelial disorder induced by chronic inflammation and oxidative stress [4, 5]. However, high-sensitivity C-reactive protein (hsCRP), the representative blood marker of inflammation, is insufficient as an index of chronic inflammation of arteriosclerotic disease because it is only an acute inflammatory marker, and its measured values vary substantially. No new inflammation-related clinical index has been established that can accurately identify the pathology of cognitive decline accompanied by DWMLs [6].

Regarding DWMLs, it has been reported that impaired blood-brain barrier function with aging promotes inflammatory cytokine production, which subsequently increases amyloid-β production [7, 8]. Inversely, a vicious cycle has been reported in which increases in amyloid-β and tau protein leading to cognitive decline were affected by chronic inflammation at the brain nerve cell level [9].

We have recently reported that adrenomedullin (ADM) plays a crucial role in the pathogenesis of DWMLs [10, 11]. ADM, which is secreted directly from vascular cells in reduced cerebral circulation-associated damage of the DWM, has a cerebral protective effect that directly inhibits the development of DWMLs [10, 11]. In addition, ADM has been shown to prevent inflammatory tissue damage in a mouse study and to exhibit brain tissue-protective effects, including vasodilator action and the inhibition of nerve cell apoptosis [11, 12]. ADM is a vasoactive peptide that exhibits a stronganti-inflammatory effect primarily produced by vascular endothelium. ADM, which can more easily cross the blood-brain barrier as a result of aging, exists as a stable fragment of an ADM precursor, mid-regional proADM (MR-proADM), in the blood. However, to our knowledge, no previous studies have investigated the significance of MR-proADM levels and DWMLs in association with cognitive decline in humans.

Therefore, the objective of this study was to investigate whether MR-proADM levels and DWMLs are associated with cognitive decline in humans.

MATERIALS AND METHODS

RESULTS

As shown in Table 1, hypertension, high baPWV, and hsCRP levels increased in accordance with DWML grade progression. A significant decrease in word fluency (animals and vegetables) was observed with DWML grade progression on the cognitive function tests. MMSE and other cognitive function scores also tended to decrease significantly with DWML grade progression (Table 1).

As shown in Fig. 2, MR-proADM levels showed a significant increase (p < 0.05) with DWML grade progression as follows: G0 (n = 72), 705.7  pg/mL; G1 (n = 137), 825.9; G2 (n = 62), 904.8; and G3 (n = 17), 1115.2. These results indicated that MR-proADM levels significantly increased in the group with severe DWML findings on MRI accompanied by cognitive dysfunction.

To investigate how closely a high MR-proADM level was correlated with DWML grade on brain MRI, the participants were divided into groups classified as DWML G2 or G3, with G0 as a reference, and binominal logistic regression analysis was performed. When 90% or higher (810  pg/mL) of MR-proADM levels in the G0 group was defined as high, as shown in Table 2, the ORs of the DWML G2 and G3 for high MR-proADM were significant: 2.27 (95% CI: 1.43–6.95) and 3.48 (95% CI: 1.88–7.16), respectively, after adjusting for age and sex, and 1.52 (95% CI: 1.29–4.05) and 3.08 (95% CI: 1.49–5.17), respectively, after adjusting for age, sex, hypertension, hyperlipidemia, diabetes mellitus, cardiovascular disease, alcohol drinking, smoking, hsCRP, education, APOE ɛ4 allele carrier, and PVH.

Interestingly, when MR-proADM levels were additionally measured in stored samples collected in the baseline survey (2008) in the G3 group, and differences with those in the second survey (2013) were investigated, a significant increase was observed, as shown in Fig. 3 (p < 0.05).

Regarding the correlation between MR-proADM levels and cognitive functional scores (Fig. 4), a significant inverse correlation was observed between word fluency test scores (animals and vegetables) and DWML grade progression (p < 0.05). Letter fluency test scores (“Ta”) also tended to decrease with increases in MR-proADM levels, but no statistically significant correlation was observed between MR-proADM levels and letter fluency (“Ka”); however, the increases in MR-proADM levels tended to be associated with lower scores for letter fluency (“Ka”) (data not shown). Moreover, an inverse correlation was found between increases in MR-proADM levels and aggravation of cognitive function scores (word recall task), indicating that blood MR-proADM levels increase in individuals with DWMLs accompanied by decreased cognitive function; this finding also suggested the usefulness of MR-proADM as a marker for early screening. PVH on brain MRI was also investigated, but in contrast to DWML grade progression, no particular tendency was noted with regard to increases in MR-proADM levels.

DISCUSSION

Cerebral arteriosclerosis accounts for a substantial percentage of dementia cases, among which, vascular dementia has been reported to account for about one-third of the total [28]. Although studies on clinical risk factors for vascular dementia have progressed relatively well, early clinical diagnosis remains difficult, and as a result, a great deal of variation has been observed in the results of epidemiological studies.

DWMLs, which reflect cerebral arteriosclerosis, are a direct risk factor for cognitive decline [3, 25], and inflammatory cytokines and chronic inflammation, which promote cerebral arteriosclerosis, are more prevalent in patients with DWMLs accompanied by cognitive decline [4, 29]. Cognitive decline is directly associated with DWML grade; therefore, DWML grade is a key factor in regard to the pathology and prevention of dementia [30].

ADM has attracted attention as a biomarker in the brain research field. ADM administration has been reported to ameliorate ischemic white matter lesions and improve cognitive function in an experimental mouse model of chronic cerebral hypoperfusion[10, 11], suggesting that ADM plays a specific role in the development and control of DWMLs. However, reliably measuring ADM remains difficult because it has a short life (turnover) in human blood and is rapidly removed from circulation. Therefore, MR-proADM, which is stable in the body and relatively easy to measure, has recently been attracting increasing attention [26]. In a large-scale cohort study recently performed in Finland, once cerebral infarction occurred, the mortality rate was higher in the group with than without high MR-proADMlevels [31].

Therefore, in this study, we investigated MR-proADM levels in groups with various grades of DWMLs using samples and brain MRI scans collected during medical checkups, and analyzed the association of MR-proADM levels with DWMLs, as well as the association between increases in MR-proADM levels with DWMLs accompanied by cognitive decline.

As shown in Table 1 and Fig. 2, MR-proADM levels significantly increased with DWML grade progression and decreases in word fluency test scores. Interestingly, as shown in Fig. 4, a direct inverse correlation was observed between increased MR-proADM levels and decreases in word and letter fluency test scores (word recall task), which suggested that MR-proADM levels could be an effective early diagnostic surrogate marker of cognitive decline.

It has been reported that increases in MR-proADM levels are a result of its production mainly by vascular endothelium and smooth muscle cells in compensatory response to cerebral hypoperfusion with cognitive decline, and this enhancement can be confirmed in peripheral blood due to hypofunction and impairment of the blood-brain barrier [12, 32]. Reimer et al. [32] performed comprehensive analysis of more than 130 candidate factors related to the development of DWMLs and observed that ADM markedly increased at a high rate in mice with reduced circulation in the brain (OR = 6.72). In our study, strong correlations were found between MR-proADM levels with DWMLs on brain MRI and cognitive decline. These results suggest that MR-proADM is useful as a surrogate marker for accurately predicting the underlying pathology of DWMLs accompanied by cognitive decline. We also investigated the association between the presence or absence of the APOE ɛ4 allele and variation in MR-proADM levels, but as shown in Table 1, no correlation was found.

Regarding the reason why MR-proADM levels significantly increased with DWML grade progression and were inversely correlated with cognitive function test scores, we propose two possibilities; these also explain our obtained data regarding changes in MR-proADM levels in accordance with the advanced grading of DWMLs with cognitive decline. The first possibility is that advanced DWMLs cause an increase in MR-proADM because it is secreted from vascular endothelium and smooth muscle cells. The other possibility is that MR-proADM levels were increased as a result of a compensatory response to the damage sustained from chronic cerebral hypofunction. This interpretation implies that MR-proADM levels increased in response to the worsening of DWMLs as a compensatory and white matter protective response.

In the present study, we did not verify which of these possibilities was most likely; however, in a previous study [11], Ihara et al. verified these explanations in a mouse model of chronic cerebral hypoperfusion after bilateral common carotid artery stenosis. They found that ADM transgenic mice with adrenomedullin-induced neovascularization (mice overexpressing circulating ADM) showed significantly faster cerebral perfusion recovery than normal mice. They also reported that working memory deficits induced by bilateral common carotid artery stenosis in mice overexpressing circulating ADM were subsequently restored. As shown in Fig. 3, MR-proADM levels significantly increased from 2008 to 2013 in direct or indirect response to worsening DWMLs based on MRI evaluation. Although we cannot say whether these increases were a result of the same responses in the human subjects, we speculated that MR-proADM levels increased in response to the damage sustained from chronic cerebral hypofunction and the worsening of DWMLs.

Based on a previous study on ADM, the vascular structure, including cerebral blood vessels, of ADM-knockout mice was immature, suggesting that ADM is essential for angiogenesis and vascular growth [33]. Inversely, blood pressure and circulatory dynamics were stable, and the incidence of arteriosclerosis and heart and kidney disorders decreased in mice overexpressing ADM, which suggested that ADM has multifaceted anti-arteriosclerosis and organ-protective actions [34–38]. Taking all of these findings into consideration, ADM is assumed to be a brain-related neuropeptide induced to improve reduced cerebral circulation in impaired DWM. In fact, it has recently been reported that nerve and glia cells also secrete ADM, that administered proADM has brain-protective actions, and that ventricular ADM administration protects brain function and inhibits excess salt ingestion and poor eating behaviors [39, 40]. In contrast to ADM, which is rapidly degraded in the blood and not considered a promising biomarker, MR-pro-ADM is a stable fragment of pre-proADM that has potential as a biomarker that can reflect the degree of local hypoperfusion. As previously reported [41], changes in MR-pro-ADM levels can be considered parallel to changes in ADM levels in a one-to-one relationship; therefore, MR-pro-ADM levels can be directly translated into ADM levels, which enabled us to confirm the candidacy of MR-pro-ADM as a new potential surrogate marker to reflect DWMLs on brain MRI. MR-ProADM is a novel predictive factor for vascular cognitive decline, and is therefore expected to provide new knowledge to the medical field. We are planning to perform a long-term follow-up study with the same participants of the current study to investigate these possibilities, and we are currently preparing a system to follow changes in blood MR-proADM levels, brain MRI findings, and cognitive function test scores over the next 5 years. In addition, Hampel et al. [42] reported finding a strong correlation between ADM and atrial natriuretic peptide (ANP) in mild cognitive impairment (MCI) cases. Although we could not measure ANP in the present study because of limited funding and samples, it will be important to add an analysis of ANP in relation to the function of ADM in the future. Therefore, we plan to measure both MR-proADM and ANP at the same time in a 5-year follow-up survey.

The present study had several major limitations. First, it had a relatively small sample size, which resulted in insufficient statistical power to assess the study hypothesis with a high degree of confidence. To evaluate the diagnostic value of MR-proADM, a large-scale cohort study is needed. Second, the data were obtained from a single center enrolling consecutive subjects who responded to a random direct mail recruitment approach in order to minimize the effect of individual selection bias. Although random mailing was used for recruitment, a small degree of selection bias was unavoidable. In summary, the small sample size and individual selection bias, as well as the limited number of DWML events and the relatively short follow-up period, meant that this study had insufficient statistical power to detect significant differences in the analysis of DWML events. Third, as shown in Fig. 4, we confirmed that the distribution of MR-proADM levels is widely spread. This variation is unavoidable because it was a human study involving sampling at only two points. We did not perform blood sampling during the baseline or second survey because of limited research funding and labor. As mentioned above, to confirm that MR-proADM levels are a good surrogate marker, we are planning to perform a long-term follow-up study of blood MR-proADM levels with the same participants of the present study. The fourth limitation was the study design. Initially, we hoped to enroll only subjects diagnosed as having vascular mild cognitive impairment. However, we could not diagnose the study subjects based on scores on the Clinical Dementia Rating (CDR) scales [43] or the cognitive subscale of the Alzheimer’s Disease Assessment Scale (ADAS-Cog) [44], which are widely recognized as being useful in diagnosing MCI, because of limited time and manpower. Therefore, the subjects were diagnosed based on routine medical checkups. To avoid exaggerating the results by definitely diagnosing subjects as having MCI, we chose not to strictly define MCI as a diagnosis in this study. However, in a follow-up survey, we plan to assess the subjects using CDR or ADAS-Cog scores.

CONCLUSION

Increases in MR-proADM levels were correlated with the grade progression of DWMLs accompanied by cognitive decline. This finding suggests the potential usefulness of MR-proADM as a clinical index of vascular cognitive disorder. Although the cause and pathology of cognitive decline with DWMLs have varied greatly, the MR-proADM level may be an effective candidate as a potential diagnostic surrogate marker of cognitive decline. Additional clinical studies performed at routine clinical practice sites and epidemiological studies involving resident cohorts are warranted to confirm the findings of the present study. Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/16-0901r1).