Serum MIF, IL-6, IL-18 and IFN-γ levels in sarcoidosis patients. We enrolled 70 subjects in this study comprising 55 sarcoidosis patients and 15 heathy subjects. Diagnosis of sarcoidosis was established based on American Thoracic Society criteria and exclusion of infection by examining samples for bacterial and viral infection[16]. All sarcoidosis subjects were ambulatory patients and sera were collected during first clinical encounter, when they were on no prior treatment. Clinical characteristics of the study subjects are demonstrated in Table 1.
Table 1
Characteristic | Patient Subjects n=55 | Control Subjects n=15 | p-value |
Age, Years (mean ± SD) | 48 ± 11 | 40 ± 10 | 0.47 |
Sex (female) | 34 (61) | 10 (66) | 0.91 |
BMI, kg.m−2 (mean ± SD) | 30 ± 7 | 27 ± 5 | 0.38 |
Race (n, %) | | | |
African - American | 49 (89) | 12 (80) | 0.52 |
White | 3 (5) | 3(20) | 0.31 |
Other | 3 (5) | 0 | 0.08 |
Smoker (PY)(mean ± SD) | 5 ± 8 | 0 | 0.00 |
Organ Involvements (n, %) | | |
Pulmonary Involvement | 54 (98) | NA | -- |
Extra Pulmonary Involvement | 42 (76) | NA | -- |
Skin | 28 (50) | NA | -- |
Eye | 24 (43) | NA | -- |
Heart | 5 (9) | NA | -- |
CNS a | 4 (7) | NA | -- |
Other | 20 (36) | NA | -- |
Initial Chest Radiograph Stages (n, %) | | | |
Stage 0 | 2 (3) | NA | -- |
Stage 1 | 9 (16) | NA | -- |
Stage 2 | 31 (56) | NA | -- |
Stage 3 | 6 (10) | NA | -- |
Stage 4 | 4 (7) | NA | -- |
Initial PFT b, Mean (IQR c) | | | |
FVC d (% predicted) | 86 (44 - 123) | NA | -- |
FEV1 e (% predicted) | 80 (27 - 121) | NA | -- |
FEV1/FVC (% predicted) | 74 (36 - 89) | NA | -- |
TLC f (% predicted) | 80 (37 - 116) | NA | -- |
DL,CO g (% predicted) | 65 (27 - 99) | NA | -- |
a Central nerve system, b Pulmonary function test, c Interquartile range, d Forced vital capacity, e Forced expiratory volume, f Total lung capacity, g Diffusing capacity of the lung for carbon monoxide |
Production of cytokines is a well-regulated host defense mechanism against pathogens that is modulated by cellular cues and the cytokine milieu. To understand the regulation of group of cytokines implicated in sarcoidosis, we assessed the levels of IL-6, IL-18, IFN-γ and MIF in the serum samples of sarcoidosis patients (n=55) and healthy controls (n=15) via ELISA. IL-6 is a pro-inflammatory cytokine and has been suggested to play a role in the pathogenesis of sarcoidosis [19]. IL-18 belongs to IL-1 family of cytokines produced by monocytes and macrophages that has been shown to be increased in sarcoidosis patients [20, 21]. Among 55 sarcoidosis subjects, only 10 patients show measurable IL-6 levels, and there was no significant difference (p= 0.47) in the levels of IL-6 between patients and healthy controls (Figure 1A). In contrast, IL-18 was detectable in the sera from all participants (Figure 1B). Sarcoidosis patients expressed significantly higher levels of IL-18 (mean± SD: 274±202 pg/mL) as compared to healthy controls (53±69 pg/mL). IFN-γ levels were also higher (71±145 pg/mL) in sarcoidosis as compared to the healthy group (18±40 pg/mL) (Figure 1C). Similarly, we assessed the MIF in the serum of our study subjects. Surprisingly, all healthy subjects had detectable MIF levels (237±219 pg/mL). The Figure 1D shows that the sarcoidosis patients have significantly higher levels of MIF. The mean values of MIF levels in sarcoidosis patients were 519±809 pg/mL. As the Box plot indicates, there were a wide variation among sarcoidosis patients in regard to MIF levels (Fig. 1D). In fact, MIF was not measurable in a large number of patients or if it was detected, the levels were below the MIF mean values of healthy subjects. In contrast, some patients exhibited very high MIF levels. To further gain insight into the potential clinical value of MIF, we classified our sarcoidosis cohort based on MIF values into two groups and performed subgroup analysis. The mean serum MIF-value in the healthy group (237pg/mL), therefore, this value was selected as the cut off value. The serum MIF values of less than 237pg/mL were classified as low serum MIF-group, whereas MIF higher than 237pg/mL were classified as high MIF-group (Figure 2A). First, we analyzed the characteristics of these two groups in terms of clinical features and other clinical biomarkers. Clinical characteristics of patients stratified by low and high serum MIF levels are provided in Table 2. We found that the low serum MIF group was characterized by greater multiorgan involvements including, cardiac (p=0.05) and CNS (p=0.09) as compared to the high MIF group. Interestingly, patients with extensive skin disease exhibited elevated serum MIF levels (high MIF group).
Table 2
Clinical Characteristics Based on Serum MIF Level
Clinical Characteristic | Low MIF Serum n=32(58%) | High MIF Serum n=23(42%) | p-value |
Smoker (PY)(mean ± SD) | 4 ± 8 | 2.5 ± 4 | 0.32 |
Serum Indicator (mean ± SD) | | | |
MIF a (pg/mL) | 90 ± 81 | 1218 ± 1058 | 0.001 |
Interferon-γ (pg/mL) | 23 ± 21 | 135 ± 36 | 0.01 |
Interleukin-10 (pg/mL) | 0 | 833 ± 1494 | 0.03 |
Interleukin-18 (pg/mL) | 264 ± 206 | 290 ± 200 | 0.90 |
Lysozyme (pg/mL) | 25 ± 1 | 21 ± 1 | 0.03 |
CRP b (mg/L) | 22 ± 10 | 9 ± 2 | 0.22 |
IgG c (mg/dL) | 1613 ± 575 | 999 ± 468 | 0.03 |
Organ Involvements (n, %) | | | |
Pulmonary Involvement | 32 (100) | 22 (95) | 0.20 |
Extra Pulmonary Involvement | 26 (81) | 12 (52) | 0.05 |
Skin | 18 (56) | 10 (43) | 0.30 |
Eye | 15 (46) | 9 (39) | 0.71 |
Heart | 5 (15) | 0 | 0.05 |
CNS d | 4 (12) | 0 | 0.09 |
Other | 12 (37) | 8 (34) | 0.47 |
Initial PFT e (mean, IQR f) | | | |
FVC g (% predicted) | 81 (60 - 105) | 82 (44 - 102) | 0.70 |
FEV1h (% predicted) | 73 (27 - 106) | 72 (44 - 98) | 0.91 |
FEV1/FVC (% predicted) | 67 (36 - 85) | 71 (43 - 88) | 0.24 |
TLC K (% predicted) | 81 (54 - 96) | 79 (37 - 102) | 0.98 |
DLCO s (% predicted) | 57 (50 - 71) | 64 (34 - 97) | 0.40 |
Follow Up PFT (mean, IQR) | | | |
FVC (% predicted) | 92 (37 - 123) | 95 (51 - 134) | 0.72 |
FEV1 (% predicted) | 82 (35 - 112) | 82 (31 - 135) | 0.95 |
FEV1/FVC (% predicted) | 74 (36 - 86) | 67 (39 - 80) | 0.11 |
TLC (% predicted) | 82 (45 - 114) | 86 (73 - 107) | 0.55 |
DLCO (% predicted) | 62 (31 - 94) | 71 (37 - 103) | 0.16 |
a Macrophage inhibitory factor, b C-Reactive protein, c Immunoglobulin G, d Central nerve system, e Pulmonary function test, f Interquartile range, g Forced vital capacity, h Forced expiratory volume, K Total lung capacity, s Diffusing capacity of lung for carbon monoxide |
Sarcoidosis patients with low MIF levels exhibit elevated lysozymes but decreased IFN-γ and IL-10. MIF has important immunomodulatory functions, both protective and detrimental roles, in the innate and adaptive immune system[2, 8, 15]. To understand the relationship between low and high levels of MIF with levels of cytokines and other parameters, we examined levels of CRP, lysozyme stratified based on MIF levels. We observed a trend in the low MIF group to have an elevated CRP (Figure 2B), but the difference did not reach statistical significance. Assessment of serum lysozyme is routinely used for evaluation of sarcoidosis [22, 23] and it is considered to be a marker for macrophage and monocytic activation [24]. Similarly, hypergammaglobulinemia is a feature of sarcoidosis and elevated levels of IgG has been reported [25]. Classification based on MIF values showed that patients with low MIF levels exhibit higher lysozyme levels (p=0.03) (Figure 2C) and elevated IgG levels (p=0.03) (Figure 2D). IFN-γ is a pleotropic cytokine modulating both the innate and adaptive immune response against pathogens and plays a role in sarcoidosis [26]. We found that higher IFN-γ levels were almost exclusively observed in the high MIF group compared to the low MIF group (p=0.01) (Figure 2E). Next, we examined the levels of IL-10 stratified by MIF levels. IL-10 has potent anti-inflammatory activities by suppressing the granuloma formation [27]. Figure 2F shows a significant difference (p=0.03) between two MIF group in terms of IL-10 levels.
Correlation of serum cytokines, lysozyme and IgG. To gain additional insights in the pathophysiological roles of measured mediators and their correlation, we performed a two-tailed Pearson correlation (Table 3). We found no significant correlation between serum MIF and IL-18. In contrast, there were a significant correlation between MIF and IFN-γ (r=0.57; p=0.001) (Figure 3A). Surprisingly, we found that MIF values highly correlates with IL-10 values (r=0.82; p value< 0.001) (Figure 2B). Similarly, we determined the correlation between serum immunoglobulin G (IgG) and MIF and found that serum MIF negatively correlates (r=-0.55, p=0.01) with IgG levels (Figure 3C). Serum lysozyme significantly positively correlated to IL-18 (r=0.59, p=0.001) but negatively to IFN-γ (r=-0.3; p=0.03) (Figure 3D and 3E). Although serum lysozyme and MIF correlated negatively but did not reach statistical significance (p=0.16). There was a significant correlation between IFN-γ and IL-10 (r=0.57; p<0.001).
Table 3
Serum Cytokines and Biomarkers Correlations
| MIF | IFN-γ | IL-10 | IL-18 | Lysozyme | CRP | IgG |
MIF | Correlation | 1 | 0.57** | 0.82** | 0.05 | -0.20 | -0.17 | -0.55** |
Sig. | | 0.001 | <0.001 | 0.71 | 0.16 | 0.31 | 0.01 |
IFN-γ | Correlation | 0.57** | 1 | 0.57** | -0.21 | -0.30* | -0.14 | -0.43 |
Sig. | 0.001 | | <0.001 | 0.15 | 0.03 | 0.39 | 0.07 |
IL-10 | Correlation | 0.82** | 0.57** | 1 | -0.08 | -0.21 | 0.35 | -0.62* |
Sig. | <0.001 | <0.001 | | 0.60 | 0.21 | 0.23 | 0.03 |
IL-18 | Correlation | 0.05 | -0.21 | -0.08 | 1 | 0.59** | 0.40* | 0.40 |
Sig. | 0.71 | 0.15 | 0.60 | | 0.001 | 0.01 | 0.07 |
Lysozyme | Correlation | -0.20 | -0.30* | -0.21 | 0.59** | 1 | 0.32* | 0.14 |
Sig. | 0.16 | 0.03 | 0.21 | 0.001 | | 0.05 | 0.54 |
CRP | Correlation | -0.17 | -0.14 | 0.35 | 0.40* | 0.32* | 1 | 0.30 |
Sig. | 0.31 | 0.39 | 0.23 | 0.01 | 0.05 | | 0.23 |
IgG | Correlation | -0.55** | -0.43 | -0.62* | 0.40 | 0.14 | 0.30 | 1 |
Sig. | 0.01 | 0.07 | 0.03 | 0.07 | 0.54 | 0.23 | |
**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed). |
Correlation of serum MIF and IL-18 with %TLC and %DLCO. Longitudinal measurements of lung function, such as % predicted total lung capacity (%TLC) and diffusion capacity of carbon monoxide (%DLCO) provide invaluable information about the progression of pulmonary sarcoidosis [28]. We evaluated the relationship of cytokine levels with PFT values: 1) PFTs at the time of sarcoidosis diagnosis and 2) PFTs obtained after 3 years follow-up. We found no correlation between the cytokines and first PFT values. Next, we evaluated Pearson correlation using %TLC values after 3 years follow up. Figure 4A shows that there is a positive, but not significant association between MIF and %TLC (r=0.14, p=0.45). However, MIF shows a significant correlation with %DLCO (r=0.4, p=0.02) (Figure 4B). There was a negative correlation of IL-18 with %TLC (r=-0.36, p=0.03) (Figure 4C). In contrast, IL-18 value tended to have negative association with %DLCO (r=-0.12) but lack the statistical significance (p=0.48) (Figure 4D). None of other markers, including lysozyme, CRP showed any correlation with PFT values.
MIF, IL-18, IFN-γ, IL-6 and IL-10 levels in BAL of sarcoidosis patients. Sarcoidosis is a systemic disease with prominent lung involvement, but the degree of lung involvement and its progression varies among patients. It has been recognized that serum cytokine level may not mirror the cytokine values of the lungs or other tissues [29]. Next, we asked if these cytokines can be measured in BAL samples. All BAL and serum samples were collected, when the first diagnosis of sarcoidosis was established, and none were receiving any immune modulatory agents. The levels of MIF, IL-18, IFN-γ, IL-6 and IL-10 were measured in the BAL samples. About two third of patients had a detectable MIF in their BAL-samples with the mean value of 822±901 pg/mL. Half of patients had a detectable IL-18 with the mean value of and 17±21 pg/mL. IFN-γ was measurable in only one BAL sample (534 pg/mL). Similarly, IL-6 was detectable only in three BAL samples with the mean value of 3.2 pg/mL. In contrast, none of patients expressed IL-10 in their BALs. Figure 5A shows the MIF values in BALs corresponding to their serum MIF values. The average of BAL MIF levels is significantly greater than the serum MIF level (p<0.001) (Figure 4A). In contrast, the average of serum IL-18 levels was significantly greater than the BAL IL-18 values (p<0.001) (Figure 4B). We found a negative correlation between MIF and IL-18 levels in BAL samples (r=-0.35, p=0.02) (Figure 4C).