CCL8 is a potential molecular candidate for the diagnosis of graft-versus-host disease

Abstract

Although graft-versus-host disease (GVHD) is a life-threatening complication of hematopoietic stem-cell transplantation (HSCT), its current diagnosis depends mainly on clinical manifestations and invasive biopsies. Specific biomarkers for GVHD would facilitate early and accurate recognition of this grave condition. Using proteomics, we screened for plasma proteins specific for GVHD in a mouse model. One peak with 8972-Da molecular mass (m/z) retained a discriminatory value in 2 diagnostic groups (GVHD and normal controls) with increased expression in the disease and decreased expression during cyclosporin A treatment, and was barely detectable in syngeneic transplantation. Purification and mass analysis identified this molecule as CCL8, a member of a large chemokine family. In human samples, the serum concentration of CCL8 correlated closely with GVHD severity. All non-GVHD samples contained less than 48 pg/mL (mean +/- SE: 22.5 +/- 5.5 pg/mL, range: 12.6-48.0 pg/mL, n = 7). In sharp contrast, CCL8 was highly up-regulated in GVHD sera ranging from 52.0 to 333.6 pg/mL (mean +/- SE: 165.0 +/- 39.8 pg/mL, n = 7). Strikingly, 2 patients with severe fatal GVHD had extremely high levels of CCL8 (333.6 and 290.4 pg/mL. CCL8 is a promising specific serum marker for the early and accurate diagnosis of GVHD.

Figure 1

Increase of 8972-Da peak in mouse GVHD plasma. (A) Representative SELDI spectra of the 8972-Da peak at day 0 obtained from pretransplantation and GVHD plasma (days 7, 14, 21, and 28 after transplantation) from an individual mouse are shown. The m/z range of 6000 to 10 000 m/z is shown. The box highlights an 8972-Da peak that is increased in intensity in GVHD plasma compared with day-0 samples. Number at the top and bottom of the figure indicates m/z. The left column indicates relative intensity of the ion peak. The Peak intensity of 8972 Da is 4.6, 47.3, 51.4, 25.2, and 55.9 on days 0, 7, 14, 21, and 28, respectively. (B) Mean normalized intensity values for 8972-Da peak in samples at each time point (day 0 is before transplantation; days 7, 14, 21, and 28 are after transplantation). The average intensity ± SE is depicted (n = 9). All GVHD samples (closed column except day-0 samples) were significantly higher than that of syngeneic controls (open column) on posttransplantation days 7, 14, 21, and 28 as judged by an unpaired t test. Increases on days 7, 14, 21, and 28 of allogeneic samples were also significantly higher than that of day-0 samples (before transplantation) (paired t test). Asterisk indicates P value.

Figure 2

Decrease of 8972-Da peak after CsA administration in GVHD mice. (A) Representative SELDI spectra of control plasma (day 0; before transplantation) and GVHD plasma (days 7, 14, 21, and 28 after transplantation) from the same mouse. The m/z range of 6000 to 10 000 m/z is shown. CsA was administered from day 8 through day 13 after transplantation, indicated by closed triangle. The box highlights a plasma peak with an average mass of 8972 Da that decreased after administration of CsA. The peak intensity of 8972 Da is 5.8, 69.4, 13.8, 15.7, and 14.3 on days 0, 7, 14, 21, and 28, respectively. (B) Mean normalized intensity for the 8972-Da peak in the samples on days 0 (before transplantation), 7, 14, 21, and 28 (n = 6, respectively). The average intensity (± SE) is depicted (n = 6). The 8972-Da peak seen at day 7 significantly decreased at days 14, 21, and 28 (paired t test). The peak at day 7 is significantly higher than that of day 0 (before transplantation) (paired t test). Asterisk indicates P value.

Figure 3

LPS or poly(I:C) treatment did not increase the 8972-Da peak. (A) At day 7 after transplantation with syngeneic BM cells (syn), mice were treated with LPS, poly(I:C), or PBS. The average value ± SE is depicted (n = 6). No effect of LPS or poly(I:C) on the 8972-Da peak was detectable, although allogeneic marrow (allo) induced an apparent increase of the 8972-Da peak. The average intensity of the peak ± SE was 4.7 ± 0.69, 5.4 ± 1.09, 4.55 ± 0.79, and 69.3 ± 7.02 on PBS, LPS, and poly(I:C) treatment of syngeneic transplantation and untreated allogeneic transplantation, respectively (n = 6). Asterisk indicates P value. (B) LPS or poly(I:C) increased serum IL-6 concentration dramatically. In a clear contrast, concentration of IL-6 was very low in allogeneic samples similar to PBS treatment (unpaired t test). The average concentration (pg/mL) of IL-6 ± SE was 1.38 ± 0.85, 32 215.5 ± 1759.3, 2087 ± 151.3, and 1.25 ± 0.13 on PBS, LPS, and poly(I:C) treatment of syngeneic transplantation and untreated allogeneic transplantation, respectively (n = 6). The difference of PBS treatment and allogeneic transplantation is not statistically significant. Asterisk indicates P value.

Figure 4

Purification of 8972-Da protein from murine plasma. Results from GVHD (A,C,E) and control (B,D,F) plasma are shown. (A,B) SELDI-TOF analyses of HPLC fraction 31.0 to 31.5 minutes derived from GVHD (A) and control (B) samples are shown. The box identifies a peak of 8972 Da that is overexpressed in GVHD (A), but not detectable in control plasma (B). (C,D) 2DE gel images of GVHD (C) and control (D) samples are shown. Plasma proteins (70 μg) of HPLC-purified fractions (pool of 31.0- to 31.5-minute fraction) from GVHD and control samples, respectively, were separated on pH 3-11 IPG in the first dimension, followed by SDS-PAGE on 8% to 20% gradient gel in the second dimension. The separated proteins were stained with silver nitrate. Insert is magnified in panels E,F. (E,F) Detailed views with apparent molecular mass of between 6500 and 14 400 Da in the 2DE gel images of GVHD (C) and control (D) samples are shown. Arrowhead indicates the candidate spot for the 8972-Da peak in GVHD (E) sample, which is not apparent in the control (F) sample.

Figure 5

Protein identification using mass spectrometry. (A) A representative MS spectrum (682.3332 m/z) of the peptide from the candidate spot in Figure 4C is shown, which was identified by nanoLC-MS/MS. (B) The deduced sequence is confirmed by MS/MS spectra of b and y ion series as CCL8 peptide (QGMSLCVDPTQK) corresponding to 682.3332 parent ion in panel A. Cysteine is carbamidomethylated in this peptide. (C) Verification of CCL8 as the 8972-Da peak by immunoSELDI is shown. The box highlights the elevation of CCL8 in the GVHD sample but not in the control sample.

Figure 6

CCL8 expression in human serum after HSCT. Serum was obtained at several time points after HSCT. The results of immunoSELDI using antibody specific for human CCL8 are shown. Box highlights peak of human CCL8. The predicted molecular mass of mature human CCL8 is 8920 Da. (A) Patient A developed clinical GVHD on day 13 after HSCT and methyl prednisolone (mPSL) was administered the same day. The peak profile of CCL8 was very weak on day 0 with an obvious increase on day 10. CCL8 then declined on day 21 after mPSL treatment, but was again up-regulated on day 33. The peak intensity of 8920 Da was 9.6, 40.6, 19.5, and 52.5 on days 0, 10, 21, and 33, respectively. (B) In patient B, no CCL8 peak was detected on day 0, but was increased on days 9 and 19, and decreased on days 21 and 30. Clinical GVHD was overt on day 19 and mPSL treatment was started. The peak intensity of 8920 Da is 2.1, 14.3, 41.4, 12.3, and 12.7 on days 0, 9, 19, 21, and 30, respectively. (C) No GVHD developed in patient C throughout the course. CCL8 peak was very low in all samples examined. The peak intensity of 8920 Da is 5.8, 12.8, 4.1, 6.5, 6.2, and 3.0 on days −7, −1, 10, 20, 34, and 38, respectively. (D) An example of CCL8 expression in a healthy control is shown. CCL8 peak is very low. The peak intensity of 8920 Da is 5.3.

Figure 7

Concentration of CCL8 in human sera. Concentration of CCL8 in human sera was determined by ELISA for human CCL8. All sera from those who did not develop GVHD after HSCT (non-GVHD/HSCT) contained less than 48 pg/mL (mean ± SE: 22.5 ± 5.5 pg/mL, range: 12.6-48.0 pg/mL, n = 7). CCL8 was highly up-regulated in GVHD plasma (GVHD/HSCT) ranging from 52.0 to 333.6 pg/mL (mean ± SE: 165.0 ± 39.8 pg/mL, n = 7). Two patients with severe, eventually fatal, GVHD showed extremely high levels of CCL8 (ie, 333.6 and 290.4 pg/mL). In healthy controls (donors), mean concentration of CCL8 was 18.9 pg/mL (range: 0.00 to 32.6 pg/mL, n = 8). The difference in serum concentration of CCL8 was statistically significant between non-GVHD and GVHD plasma obtained from patients undergoing HSCT.