Alterations to the Frequency and Function of Peripheral Blood Monocytes and Associations with Chronic Disease in the Advanced-Age, Frail Elderly

Background Circulating myeloid cells are important mediators of the inflammatory response, acting as a major source of resident tissue antigen presenting cells and serum cytokines. They represent a number of distinct subpopulations whose functional capacity and relative concentrations are known to change with age. Little is known of these changes in the very old and physically frail, a rapidly increasing proportion of the North American population. Design In the following study the frequency and receptor expression of blood monocytes and dendritic cells (DCs) were characterized in a sample of advanced-age, frail elderly (81–100 yrs), and compared against that of adults (19–59 yrs), and community-dwelling seniors (61–76 yrs). Cytokine responses following TLR stimulation were also investigated, as well as associations between immunophenotyping parameters and chronic diseases. Results The advanced-age, frail elderly had significantly fewer CD14(++) and CD14(+)CD16(+), but not CD14(++)CD16(+) monocytes, fewer plasmacytoid and myeloid DCs, and a lower frequency of monocytes expressing the chemokine receptors CCR2 and CX3CR1. At baseline and following stimulation with TLR-2 and -4 agonists, monocytes from the advanced-age, frail elderly produced more TNF than adults, although the overall induction was significantly lower. Finally, monocyte subset frequency and CX3CR1 expression was positively associated with dementia, while negatively associated with anemia and diabetes in the advanced-age, frail elderly. Conclusions These data demonstrate that blood monocyte frequency and phenotype are altered in the advanced-age, frail elderly and that these changes correlate with certain chronic diseases. Whether these changes contribute to or are caused by these conditions warrants further investigation.


Introduction
Age-related changes in circulating immune cell composition and levels of circulating pro-inflammatory cytokines have been associated with longevity [1], frailty [2] and age-related diseases such as Alzheimer's and Parkinson's disease [3], and rheumatoid arthritis [4]. Although the original description of the ''immune risk phenotype'' (a constellation of immunological markers that is predictive of survival in the aged) consisted primarily of T cell markers and levels of circulating pro-inflammatory cytokines [5], recent studies have begun to investigate age-related changes in myeloid cells such as monocytes and dendritic cells [6][7][8][9][10][11]. In the peripheral blood, monocytes can be subdivided into the classical (CD14 ++ HLA-DR + ), intermediate (CD14 ++ CD16 + HLA-DR + ) and non-classical (CD14 +/dim CD16 + HLA-DR + ) subsets, and dendritic cells can be subdivided into myeloid CD1c + HLA-DR + or CD141 ++ HLA-DR + subsets and the plasmacytoid (pDCs, CD303 + HLA-DR + or CD123 ++ HLA-DR + ) subset [12]. For individuals that are particularly susceptible to developing infectious or chronic disease, such as the advanced-age, frail elderly, alterations to these cellular populations may be a sensitive biomarker in determining their level of risk. These markers could include the frequency of a given cellular subset in the circulation, the expression of receptors that are critical for the migration to tissues via chemokine gradients or the innate response to pathogens, or potentially the ex vivo response to an exogenous stimuli.
In the following study, we characterized the frequency of blood monocytes and DCs, as well as their expression of the innate signalling receptors toll-like receptor (TLR) -2 and -4, and the chemokine receptors CCR2 and CX 3 CR1. Furthermore, we sought to test the hypothesis that monocytes from the advancedage, frail elderly are immunosenescent, and therefore are likely to be less responsive to innate ligands for TLR-2 (Pam3CSK4) and -4 (lipopolysaccharide, LPS), compared to adults. To investigate whether monocyte and DC frequency and phenotype associate with chronic diseases common to the very old, we recruited a second, larger cohort of the advanced-age, frail elderly.

Participants
Young and middle-aged adults (19-59 years old, median = 34, n = 35 (42% female)) and community-dwelling seniors (61-76 years old, median = 69, n = 45 (67% female)) were recruited from Hamilton, Ontario between January and May in 2012. The advanced-age, frail elderly (defined as having a score of at least 4 on the Clinical Frailty Scale [13]) were recruited from five local nursing homes in 2010 and 2012. Participants were excluded if they were currently on immunosuppressive medication and preexisting diseases were established by review of each participant's medical chart (Table 1). Participants recruited between January and May in 2012 (81-100 years old, median = 89, n = 49, 88% female) were compared against adults and community-dwelling seniors with regards to monocyte and DC frequency and phenotype, while participants recruited between September and December in 2010 (68-99 years old, median = 88, n = 136, 85% female) were examined for associations between those immunophenotyping variables and pre-existing diseases. The latter, second cohort was deemed necessary in order to have sufficient statistical power to perform the desired association tests. For all participants venous blood was collected from all donors by sodium heparin vacutainer (BD Biosciences, NJ, USA). Written informed consent was obtained from all participants or their legally appointed guardian in the event they were not competent to provide consent themselves. These studies and consent procedures and documents were approved by the McMaster Research Ethics Board (#13-05-14).

Statistics
All statistical analyses were performed in R 2.11.1 (R Development Core Team, 2011) or Microsoft Excel. For immunophenotyping, differences between age groups were compared using the non-parametric Wilcoxon rank sum test. Experimental-wise significance threshold was determined using the Benjamini-Hochberg procedure for controlling false discovery rate. To determine if donor sex provided substantial bias in our comparisons between age groups, we performed an initial analysis by linear regression on log-transformed values. This indicated that sex only had a significant (experimental-wise p,0.05) effect on the absolute count and proportion of classical monocytes. For associations with disease in the advanced-age, frail elderly, analysis was performed by logistic regression on log-transformed parameters, and adjusted for age. Logarithmic transformation was necessary to approximate normality, and the ratio of males to females was determined to be balanced between cases and controls (Chi-square p.0.05). Comparison of intracellular cytokine production was performed by Student's t-test on log-transformed values.

Results
For our characterization of peripheral blood monocyte and DC subsets in the advanced-age, frail elderly, we included a cohort of young adults and community-dwelling seniors in order to ascertain whether the cellular frequency and receptor expression levels observed are consistent with alterations that occur over the course of aging, or if they are indeed particular to advanced-age, frail elderly individuals. The absolute frequencies of CD45+ PBMCs were found to significantly decrease with age (mean 6 SEM: young adults 2,3396100, seniors 1,8936120 and advanced-age, frail elderly 1,1466107) ( Table 2) whereas there was a decrease in the absolute number and percentage of classical monocytes between young adults and the aged (seniors and advanced-age, frail elderly), but no significant decrease between seniors and the advanced-age, frail elderly (Table 2). Consistent with previous studies [8,10], the ratio of classical to intermediate monocytes is reduced with age and we observe that this reduction is more dramatic in the advanced-age, frail elderly (mean 6 SEM: young adults: 25.561.8, seniors: 20.562.4, advanced-age, frail elderly: 14.461.3). As has been previously observed [6,7,9], there was a reduction in circulating myeloid (CD1c+ and CD141+) DCs, which we found is further decreased in the advanced-age, frail elderly, while plasmacytoid DCs were significantly reduced in seniors and the advanced-age, frail elderly.
In addition to measurements of frequency, the expression of innate pattern recognition receptors TLR-2 and TLR-4 were measured on monocytes and DCs, and the expression of chemokine receptors CX 3 CR1 and CCR2 on monocytes alone (Table 2). It would appear that the percentage of TLR-2 expressing myeloid DCs is increased in the advanced-age, frail elderly, while no differences were observed for monocytes. It should be noted that although the trends regarding TLR-2 expression suggest an increase from young adults, to seniors, to the advanced-age, frail elderly. However, the subtlety in these alterations and degree of variation do not allow us to conclude as such. A subtle, but significant increase in the percentage of TLR-4 expressing classical monocytes was also observed in the advanced-age, frail elderly. There is an age-related reduction in the percentage of monocytes expressing CCR2, but no significant difference between community-dwelling seniors and the advancedage, frail elderly. For CX 3 CR1 however, a reduction in the percentage of expressing monocytes appears to be limited to the advanced-age, frail elderly.
Although only subtle differences in the expression of TLR-2 and -4 were observed for monocytes from the advanced-age, frail elderly, we sought to additionally characterize the functional  capacity of monocyte subsets from this age group to respond to stimulus via these receptors. Using intracellular cytokine staining, the production of IL-1b, IL-6, IL-8 and TNF by monocytes subsets in response to Pam3CSK4 (TLR-2 agonist) and LPS (TLR-4 agonist) were quantified in PBMCs from the advancedage, frail elderly and young adults. Consistent with previous literature [11,15,16] the relative production of cytokine by monocyte subsets are as follows: IL-1b, Classical = Intermediate . Non-classical; IL-6, Intermediate . Classical . Non-classical; IL-8, Classical = Intermediate . Non-classical; TNF, Intermediate $ Non-classical . Classical (Figure 2A-D). No significant differences between age-groups in the overall production of IL-1b (Figure 2A) or IL-6 ( Figure 2B) were observed. Classical monocytes from the advanced-age, frail elderly produced more IL-8 in response to LPS as compared to young adults ( Figure 2D), and for all subsets, with exception to intermediate monocytes stimulated with LPS, TNF production was greater in the advanced-age, frail elderly at baseline (PBS mock control) and in response to Pam3CSK4 or LPS ( Figure 2D). Interestingly, while the overall production of TNF was greater in the advanced-age, frail elderly, the relative induction of TNF (versus PBS mock control) was significantly lower for all subsets compared to young adults ( Figure 2E). No differences between age-groups were observed for the relative induction of IL-1b, IL-6 or IL-8 (data not shown).
To determine whether the observed alterations to monocyte and DC frequency and monocyte CCR2 and CX 3 CR1 expression in the advanced-age, frail elderly are associated with chronic disease, we analyzed a larger, second cohort of 136 participants (Table 1). This cohort second cohort was deemed necessary in order to have sufficient statistical power to perform the desired association tests. Within the advanced-age, frail elderly cohort we performed logistic regression for each of the monocyte and DC markers in a univariate manner against the presence of chronic obstructive pulmonary disease, congestive heart failure, coronary artery disease, asthma, dementia, cerebral vascular accident, diabetes mellitus, arrhythmia or anemia (Table 3). Other than a positive association between pDC frequency and dementia, no significant associations were observed for the frequencies of blood DCs. Although the senior and advanced-age, frail elderly groups had fewer monocytes expressing CCR2, there was no statistically significant association between monocyte CCR2 expression and disease (data not shown). In contrast, reductions in CX 3 CR1 expression only occurred in the advanced-age, frail elderly and individuals with elevated levels of CX 3 CR1 had a greater likelihood of having dementia, while reduced expression was associated with an increased risk of diabetes and anemia ( Figure 3A). In addition there was a significant correlation between monocyte frequency and dementia, diabetes and anemia. The likelihood of having dementia was positively associated with monocytes and the classical to intermediate monocyte ratio, whereas for diabetes and anemia, opposite trends were observed ( Figure 3B).

Discussion
Our results indicate that for many, but not all myeloid cell populations, age-related alterations tend to become more pronounced with advanced-age and frailty. These changes in circulating myeloid cell populations may reflect changes in precursor generation or emigration from the bone marrow. As an example, the recent finding in mice that a reduction of circulating pDCs stimulates myelopoiesis and increases circulating myeloid-derived suppressor cells (MDSCs) [17], whose numbers increase in the advanced-age, frail elderly [14], implies that there Table 2. Cont.    Table 3. Associations between blood monocyte and DC markers and disease in the advanced-age, frail elderly.  may be previously unappreciated feedback mechanisms between circulating DCs and the bone marrow compartment. In addition to changes in frequency, changes in phenotype and function have been shown to occur with age and it has been proposed that these phenotypic changes may contribute to ageassociated chronic disease, especially those with inflammatory etiology. There have been conflicting reports, for example, as to whether monocytes have hypo-or hyper-inflammatory responses to TLR ligands and whether these might be due to changes in TLR expression [18,19]. While we observed only a slight increase in the percentage of TLR-4 expressing classical monocytes and no changes to the expression of monocyte TLR-2, the production of TNF, and to a lesser extent IL-8, was significantly higher in monocyte subsets from the advanced-age, frail elderly, both at baseline and in response to TLR-2 and -4 stimuli. This is similar to what has been shown in previous reports [8,11], and supports the theory that constitutive over-production of cytokines by monocyte subsets may predispose elderly individuals to a higher risk of chronic disease.
Another potential contributor to the development of chronic disease in the elderly is the ability of circulating monocytes to migrate to the tissues. Both CCR2 and CX 3 CR1, receptors for the chemokines MCP-1 (CCL2) and fractalkine (CX 3 CL1), are potently involved in the migration and recruitment of monocytes in the host. Monocyte recruitment via chemokine receptors has been linked to the development of inflammatory diseases such as atherosclerosis and cancer [20,21], and an increased expression of the monocyte/macrophage chemoattractant CX 3 CL1 has been observed in cardiovascular disease and Alzheimer's disease [22,23]. Although correlations of increased CX 3 CR1 expression and dementia have not been previously demonstrated in humans, our observations are consistent with mouse models of Alzheimer's disease in which either a loss of CX 3 CR1 gene expression [24] or a reduction in signalling through CX 3 CR1 [25] results in an improved outcome. Little is known regarding monocyte phenotype or CX 3 CR1 expression and diabetes, although it has been shown that monocytes display an activated phenotype in diabetics [26,27] and the production of CX 3 CL1, likely by adipocytes, is found at higher levels in diabetics [28]. Whether monocytes expressing lower levels of CX 3 CR1 contribute to insulin resistance and diabetes in the frail elderly is not known. We also observed associations of decreased monocyte numbers and CX 3 CR1 expression with anemia. Chronic inflammation anemia is extremely common in the frail elderly [29] and is associated with elevated levels of inflammatory cytokines, especially IL-6 [30]. Since monocytes and erythrocytes share a common progenitor in the bone marrow, this association may be due to a common mechanism of suppressed myelopoeisis due to the aging or the immune status of the host.
In summary, changes in monocyte frequency, phenotype and function occur in the advanced-age, frail elderly and correlate with chronic disease. However, we do not know if these changes predispose individuals to age-related diseases or whether the overall immune status associated with many of these conditions is what ultimately impacts monocyte development and function. Future longitudinal studies will need to be performed to dissect the cause and effect of these changes as individuals approach advanced-age. Changes in circulating monocyte frequency and phenotype may be robust markers of immune risk in the aged.