Study population
A cohort of 402 adult kidney graft recipients (≥17 years), transplanted in 2000 and 2001 at Oslo University Hospital Rikshospitalet, was included in the original study previously described in detail [
10]. Blood samples at the time of transplantation were available in 382 of the patients, who were included in the present study. In the present study the follow up period was extended until December 31, 2014. Data on patient survival was obtained from the Norwegian Renal Registry. Prophylaxis with trimethoprim-sulfamethoxazole against
Pneumocystis jiroveci was routinely used for 6 months after transplantation. No patients received prophylaxis against cytomegalovirus (CMV) but were treated with valganciclovir at first positive CMV antigen test.
Biochemical assays
MASP-
2. The MASP-2 assay was previously described in detail [
12]. Microtiter wells were coated with 0.5 μg anti-MASP-2 antibody (MAb clone 8B5) in 100 μl PBS overnight at 4 °C. The wells were blocked and washed with buffer. The samples were diluted 75-fold in 1 M NaCl, 10 mM Tris–HCl, 10 mM EDTA, 15 mM NaN
3, 0.05 % (
v/v) Tween 20, pH 7.4, containing 0.01 % (
w/v) heat aggregated human IgG (Beriglobin, incubated 30 min at 63 °C and centrifuged 10 min at 3000 g to remove large aggregates). The heat aggregated human IgG is included to inhibit the influence of possible rheumatoid factors in sandwich type immuno assays. A pool of plasma was used as a standard and three different plasma samples were used as internal controls and included on each microtiter plate used. Following incubation overnight at 4 °C, the wells were washed and incubated for 1.5 h at room temperature with 0.1 μg biotinylated anti-MASP-2 antibody (MAb 6G12), in 100 μl of TBS/Tw/CaCl
2 (10 mM Tris–HCl, 145 mM NaCl, 5 mM CaCl
2, 15 mM NaN
3, 0.05 % (
v/v) Tween 20, pH 7.4) containing 0.01 % (
w/v) heat aggregated human IgG and 1 % (
v/v) bovine serum. The wells were washed followed by incubation with 10 ng europium-labelled streptavidin (Perkin Elmer) in 100 μl of TBS/Tw, 25 μM EDTA for 1 h at room temperature. After wash bound europium was detected by time-resolved fluorometry after the addition of an enhancement solution (Perkin Elmer).
MASP-3. The MASP-3 assay we used was described in details by Degn et al. [
9]. Microtiter wells were coated with 0.2 μg antibody reacting with MASP-3 (MAb 5F5) in 100 μl PBS overnight at 4 °C. Residual bindings sites in the wells were blocked with HSA (1 mg/ml TBS), and the wells next received the samples diluted 100-fold in Binding buffer (20 mM Tris–HCl, 1 M NaCl, 5 mM CaCl
2, 1 mg HSA/ml, 15 mM NaN
3, 0.05 % (
v/v) Triton X-100) containing 0.01 % (
w/v) heat aggregated human IgG. A standard curve was made from a pool of citrate plasma from donor blood. The plasma pool was diluted 1/10 followed by serial 3-fold dilutions in Binding buffer. For quality control, each microtiter plate in addition received the same three citrate plasma samples diluted 100-fold. All samples, standards and controls, were tested in duplicate. Following incubation overnight at 4 °C, the wells were washed thrice with TBS/Tw/CaCl
2 and incubated for 2 h at room temperature with 25 ng biotinylated anti-MASP-3 antibody (MAb 38.12.3) in 100 μl of TBS/Tw/CaCl
2 containing 1 % (
v/v) bovine serum. The wells were subsequently washed thrice and then incubated for 1 h with europium-labeled streptavidin (Perkin Elmer) diluted 1000-fold in TBS/Tw, 25 μM EDTA. After washing with TBS/Tw/CaCl
2 and the addition of enhancement buffer the wells were read by time-resolved fluorometry.
MAp44. Levels of MAp44 were determined as previously described [
13]. Wells of microtiter plates were coated with 0.5 μg mouse anti-human MAp44 antibody (MAb 2D5) in 100 μl PBS. Wells were subsequently blocked with TBS/Tw. Serum samples, the standard and three quality controls were diluted 40-fold in Binding buffer containing 100 μg/ml of each of heat-aggregated humane IgG, bovine IgG, rat IgG and mouse IgG at pH 7.4. To construct a standard curve, standard citrate plasma samples with known MAp44 concentrations were diluted 1/10 and then a further 7 times two-fold. Samples of 100 μl were added to wells and incubated overnight at 4 °C. After incubation and washing, the wells were incubated with a biotinylated antibody reacting with MAp44 (MAb 4H2). After washing, 1000-fold diluted europium-labeled streptavidin was added; then, after incubation and washing, enhancement buffer was added. The released europium was measured by time-resolved fluorometry. Inter-assay reproducibility was assessed by determining MAp44 in three different control citrate plasma samples.
MAp19. The assay for MAp19 was previously described in detail [
14]. Wells were coated with 4 μg anti-MAp19 antibody (MAb 6G12) per ml sodium acetate buffer (50 mM Na-acetate, 145 mM NaCl, pH 4.5) o.n. at 4 °C. The wells were then blocked with HSA, 1 mg/ml TBS, washed thrice with TBS/Tw, and incubated o.n. at m temperature with serum samples diluted 20-fold in MAp19 buffer (10 mM Tris–HCl, 1 M NaCl, 10 mM EDTA, 0.05 % Tween 20) containing 100 μg heat-aggregated hIgG per ml, and 100 μg normal rat IgG per ml. After washing the wells, biotinylated antibody reacting with MAp19 (MAb 4D12) was added at 1 μg per ml TBS/Tw containing 1 mg HSA/ml. Following another wash, the wells were developed with europium-streptavidin as described above. A standard curve was prepared by applying a 2-fold serial dilution of a standard serum (8 dilutions) and a buffer control. Along with three internal control sera (high, medium, low), this was included on every plate. All samples, standards and controls were in duplicates.
Statistical analyses
The activating and inhibitory molecules of the lectin pathway (MASP-2, MASP-3, MAp44 and MAp19) were analyzed as potential predictors for mortality. Because of the unknown pattern of the association between the effector molecules and mortality, we divided the effector molecules into quartiles, and Kaplan-Meier survival plots were made for each variable. Based on the survival plots for MAp44, a cut point was made at the 25 percentile for MAp44, and the variable was dichotomized. Consequently, we studied the effect of Low MAp44 (≤1716 ng/mL; 1st quartile) versus High MAp44 (>1716 ng/mL; 2nd–4th quartile) on kidney graft and patient survival by Cox proportional hazard regression models. A number of relevant variables known to be associated with survival were tested as predictors and potential confounders in univariable COX models: CMV infection during the first 100 days after transplantation (as time-dependent variable), induction therapy with basiliximab, recipient age per year, recipient gender, living donor, donor age per year, preemptive transplantation (previously not taken in renal replacement therapy), coronary heart disease, hypertensive nephropathy and diabetic nephropathy at the time of transplantation.
Explanatory variables with p < 0.2 in the univariable analyses were included in a multivariable Cox model for overall mortality. Similar Cox model for overall mortality was conducted for sub cohort of patients that underwent dialysis before transplantation, with dialysis vintage as one of the co-variables. In the multivariable Cox models for cardiovascular mortality and mortality due to infectious diseases in total cohort and in multivariable Cox model for overall mortality in the subsample (median age of 51.7 years or below), fewer explanatory variables were included to avoid overfitting, i.e. spurious effects due to too few events per explanatory variable. Those variables were chosen based on lowest p-value in univariable analyses.
For each Cox model, the proportional hazard assumption was checked with a test based on Schoenfeld residuals and found to be adequately met. We tested potential interactions between age and biomarkers in COX analyses. First, univariable COX analyses were performed in groups according to recipient’s age quartiles. Based on these results the cohort was stratified by median age, and subsequent analyses were done separately for patients below and above median age.
Pearson bivariate correlation coefficients were estimated to explore the relationship between effector molecules that arise from same gene by mutually exclusive splicing, e.g. between MASP-3 and MAp44 or between MASP-2 and MAp19. The Pearson chi-squared test was used to compare categorical variables, and median regression was used to compare medians.
The statistical softwares SPSS (SPSS 21) and Stata 14 (StataCorp LP, College Station, TX) were used to perform the statistical analyses.