Background: Evidence from animal models supports the hypothesis that dysregulated transforming growth factor beta(1) (TGF beta(1)) expression plays a role in chronic allograft rejection, the progression of diabetic nephropathy and fibrotic glomerulopathies. However, more evidence is required to support this hypothesis in man, and the current literature concerning blood TGF beta(1) levels in clinical studies is highly confused. We have investigated: (i) the hypothesis that the widespread practice of activating clinical samples prior to measurement of TGF beta(1) is detecting the platelet-released pool of TGF beta(1), artefactually generated on venepuncture and unrepresentative of the real circulating in vivo TGF beta(1) pool; and (ii) the effect of different immunosuppressive drugs on apparent TGF beta(1) plasma levels.
Methods: The effect of two different venepuncture procedures on plasma TGF beta(1) was compared in 10 healthy volunteers, one procedure designed to minimize platelet activation and the other representing standard venepuncture practice in a clinic situation. Blood samples from 52 renal transplant recipients on either cyclosporine or tacrolimus immunosuppression were taken by standard venepuncture to investigate the effect of immunosuppressive drugs on plasma TGF beta(1). Plasma TGF beta(1) and beta thromboglobulin were measured by ELISA.
Results: Among 10 healthy volunteers who underwent two different methods of venepuncture, eight of 10 had undetectable levels of TGF beta(1) (<100 pg/ml) under conditions that minimize platelet activation. In contrast, all 10 paired plasma samples collected by vacutainer had measurable TGF beta(1) (median 7.70 ng/ml, interquartile range 5.87-13.64 ng/ml) following acid/ urea activation. The median beta TG level (a measure of platelet degranulation) was 0.71 microg/ml (interquartile range 0.53-1.19 microg/ml) in the special collections compared with 3.39 microg/ml (interquartile range 2.27-4.33 microg/ml) in the vacutainer samples (P=0.0029). Among 52 allograft recipients there was a significantly higher mean TGF beta(1) level in plasma from patients on cyclosporine therapy compared with patients on tacrolimus (28,090+/-26,860 pg/ml vs 7173+/-10 610 pg/ml, respectively; P<0.002). Mean plasma beta TG levels were also significantly higher during cyclosporine therapy compared with tacrolimus (8.14+/-5.54 microg/ml vs 3.66+/-3.32 microg/ml, respectively; P<0.002). However, when TGF beta(1) values were corrected for the degree of platelet activation (by factoring with beta TG) there was no significant difference between TGF beta(1) levels on cyclosporine or tacrolimus (4117+/-2993 pg/microg beta TG vs 2971+/-658 pg/microg beta TG, respectively; P=0.294).
Conclusions: To avoid erroneous hypotheses concerning TGF beta(1) and perpetuating confusion in the literature over levels in health and disease, it is imperative that proper internal controls for platelet activation are used. The effects of experimental treatments and drugs on platelet biology must be rigorously controlled when attempting to measure and interpret plasma levels of TGF beta(1) in clinical practice.