Abstract
Serum sCD163 as a biomarker of adipose tissue inflammation in OSA patients: limits and perspectives http://ow.ly/kgxL30d7RnE
To the Editor:
Recently, Murphy et al. [1] demonstrated, through a challenging animal, cellular and human translational approach, that intermittent hypoxia induces a pro-inflammatory activation of adipose tissue macrophages (ATMs), promoting insulin resistance. They confirmed that intermittent hypoxia lowers insulin-mediated glucose uptake in 3T3-L1 adipocytes, and evidenced the interrelationship between inflammation and insulin resistance in the adipose tissue of mice exposed to intermittent hypoxia. They then measured the concentration of serum soluble CD163 (sCD163), an assumed pro-inflammatory biomarker reflecting macrophage activation, in obstructive sleep apnoea (OSA) patients classified according to their BMI and apnoea/hypopnoea index (AHI). This is the first time serum sCD163 has been evaluated in OSA patients, and its association with AHI is promising. With the aim of enhancing its relevance in further studies, we wish to discuss the following points.
1) Activation of adipose tissue macrophages (ATMs) towards M1 phenotype in OSA patients. Murphy et al. [1] made the assumption that circulating sCD163 levels in OSA patients reflect the intermittent hypoxia-dependent polarisation of ATMs towards a M1 phenotype. As appropriately cited, Kračmerová et al. [2] showed that circulating sCD163 concentrations are associated with both macrophage content in human adipose tissue, and insulin sensitivity. However, the assumption that serum sCD163 reflects such a polarisation based on the ADAM-17-dependent cleavage, as hypothesised by Etzerodt et al. [3], is unconvincing since only shown in HEK293 cells. Human ATMs are characterised by a high expression of CD163 which, in contrast, is weakly expressed in differentiated M1 macrophages [4]. Therefore, the question arises as to whether a link exists between the ATM/M1 difference in CD163 expression and the increase in serum sCD163 concentration and, if confirmed, to what extent nocturnal intermittent hypoxia impacts this inflammatory activation. An interesting, but technically demanding, approach would be to reproduce, both on control and OSA patients, the experiments from Zeyda et al. [4] on isolated subcutaneous ATMs, using flow cytometry for ATM/M1/M2 characterisation (including CD163) and quantifying in vitro their cytokine production under stimulating agents, in addition to sCD163 measurements both in serum and in culture supernatants. Besides the ATM/M1 ratio in CD163 expression, Zeyda et al. [4] showed that ATMs produce higher levels of tumour necrosis factor-α, interleukin (IL)-6 and macrophage inflammatory protein-1α under interferon-γ stimulation, and IL-10 under lipopolysaccharide (LPS) stimulation, compared to M1 macrophages. In contrast, under LPS stimulation, M1 macrophages produced high amounts of Rantes, a chemokine not secreted by ATMs. Accordingly, it could be assumed that high circulating Rantes levels observed in obesity hypoventilation syndrome mainly result from an ATM to M1 phenotypic change [5]. Thus, although promising, the association between serum sCD163 and AHI requires further investigation before suggesting sCD163 as a circulating biomarker of ATM activation in OSA patients.
2) Relationship between ATM-specific inflammation and insulin-resistance in OSA patients. Murphy et al. [1] supported the hypothesis of an intermittent hypoxia-dependent ATM activation through a multiple regression with sCD163 as a dependent variable, from which AHI (or oxygen desaturation index) emerged as the only predictor. Unfortunately, it seems that the authors did not include the homeostasis model assessment resistance index (HOMA-IR) as an independent variable in this model, making it difficult to appreciate if sCD163-based inflammation is related to insulin resistance in these patients. Moreover, results from univariate regressions between serum sCD163 and body mass index (BMI), waist/hip ratio, HOMA-IR or lipid parameters are lacking, making it difficult to evaluate the relationship between sCD163 and insulin resistance. Indeed, such data could confirm the recent results from Cinkajzlová et al. [6], who showed plasma sCD163 as positively associated with BMI, waist/hip ratio, blood glucose, insulin, and HOMA-IR, and negatively associated with high density lipoprotein-cholesterol in a group of subjects including controls, obese and diabetic patients. Besides these statistics, the authors could also provide serum sCD163 mean±sd for the BMI- and AHI-based subgroups, which would allow us to compare the ranges from healthy subjects, (334±25 ng·mL−1), obese (467±40 ng·mL−1) and diabetic patients (514±37 ng·mL−1) [6], or from many others pathophysiological contexts [7]. Further studies could include potential correlations of serum sCD163 with more sophisticated anthropometric measurements, such as fat mass.
3) Potential impact of the liver on the relationship between serum sCD163 and OSA severity. Finally, Murphy et al. [1] relevantly reminded that, besides adipose tissue disorders, intermittent hypoxia also unbalances hepatic metabolism, favouring steatohepatitis and insulin resistance [8]. Accordingly, an increase in liver enzymes has been shown in OSA [9, 10], but still remains controversial. Recently, alanine aminotransferase has been found as positively associated with plasma sCD163 in obese subjects [6]. Thus, although patients with OSA were monitored for liver function, it would have been interesting to present liver enzyme results, and to test their associations with the circulating levels of sCD163.
In conclusion, Murphy et al. [1] highlighted for the first time the potential interest of serum sCD163 as a pro-inflammatory biomarker associated with OSA severity. However, further investigations are needed to confirm the assumption that it reflects the impact of intermittent hypoxia on ATM polarisation towards the M1 phenotype.
Acknowledgements
The authors thank Vincent Fitzpatrick for the English rereading.
Footnotes
Conflict of interest: None declared.
- Received May 22, 2017.
- Accepted May 24, 2017.
- Copyright ©ERS 2017