The pleiotropic and fundamental roles of BMPs in both normal physiological and pathophysiological processes, indicates that the ability to measure BMP activity may have clinical use. In this paper we describe the development of a highly sensitive and specific assay to measure BMP activity from biological samples. We generated a stable cell clone "C2C12BRE" expressing a luciferase reporter under the control of the BMP response elements from the mouse Id1 promoter. When compared to other available bioassays, our assay has enhanced BMP isoform range, increased sensitivity and is capable of simultaneously measuring multiple individual BMP isoforms from the same complex biological samples.
The most widely used cell based assay to assess BMP activity employs measurement of alkaline phosphatase activity in C2C12 myoblast cells . However, this assay may lack sufficient sensitivity to measure physiological levels of BMPs, takes several days to perform and is susceptible to influence from other signalling pathways [31, 32]. Recently, other cell based assays using C3HT10T1/2 embryonic mouse cells , C2C12 and HepG2 cells  have been described. These assays are capable of measuring BMP2, BMP4, BMP6 and BMP7. We show that our assay is also capable of measuring BMP9 levels from 4–200 pM and BMP10 levels from 40–205 pM. Our assay has enhanced sensitivity for BMP4 (2 pM compared to 3 pM), BMP6 (16.5 pM compared to 40 pM) and BMP7 (42 pM compared to 125 pM). ELISA assays are commercially available for the measurement of BMP4, BMP6 and BMP7 but are currently unavailable for BMP9 and BMP10. BMP4 ELISAs are reported to measure 0.03–2 ng/ml (RnD Systems, USA) and 0.015–1 ng/ml (Raybiotech, USA) similar to the sensitivity of our assay (0.05–1 ng/ml). An ELISA for BMP6 is reported to measure 0.08–8 ng/ml (Raybiotech, USA) similar to our assay sensitivity range (0.5–10 ng/ml). A BMP7 ELISA is reported to have enhanced sensitivity but limited linearity (0.01–1 ng/ml, Raybiotech, USA) when compared to our bioassay (1–10 ng/ml). Our assay therefore has similar sensitivity to ELISA assays but has the advantage that it measures biologically active BMP and not total BMP levels. Importantly, we develop the bioassay methodology further and show that by using BMP isoform specific blocking antibodies it is possible to measure the specific activity of individual BMPs in complex biological samples whereas previous methods have so far only been capable of detecting total BMP activity. Furthermore using recombinant BMPs and isoform specific blocking antibodies we demonstrate that is possible to simultaneously quantitate total BMP activity and individual isoform BMP activity from the same biological sample.
As shown in Figure 1 our assay is specific for BMPs as related TGF-β superfamily members TGF-β, Nodal and MIS had no effect on luciferase activity. By careful titration analysis we were able to define the linear ranges of activity for different BMP isoforms. Our assay is most sensitive for measuring the activity of BMP4, BMP6 and BMP9 and is capable of measuring as little as 40 pM of BMP7 and BMP10 (Figure 2). As we could measure BMP4, BMP6 and BMP9 over greater than 10 fold ranges of concentrations coupled with the availability of isoform specific blocking antibodies for these species we focused our efforts on developing the assay using these BMPs. Using recombinant BMPs and blocking antibodies we demonstrated that it is possible to determine the specific activity of individual BMPs (Figure 3).
We next used FCS as a test complex biological fluid for determining the versatility of our assay. By performing titration analysis with recombinant BMP9 and serum stimulation of C2C12BRE cells with and without BMP9 blocking antibody we discovered that FCS contains high levels of biologically active BMP9 ranging from 5–10 ng/ml (Figures 4, 5, 6). Whilst this paper was in preparation, Sabine Bailly and colleagues discovered that BMP9 is present in human serum at similar concentrations from 2–12 ng/ml and acts as a circulating vascular quiescence factor . Our data indicates that BMP9 may also play roles during bovine development and reveals a remarkably conserved steady state level of BMP9 in serum samples between species. BMP9 has been shown to signal via ALK1 in endothelial cells  and indicates that either C2C12 cells express functional ALK1 or that BMP9 may also engage alternative receptors. We are currently investigating these hypotheses. Our analysis also revealed that FCS contains BMP4 ranging from 3–5 ng/ml and BMP6 ranging from 2.75–8 ng/ml (Figure 6). BMP4 has previously been purified from FCS  but our study is the first to measure the concentration of BMP4 in FCS. To our knowledge this current study is the first to identify BMP6 in FCS. Our experiments also revealed that in two out of three FCS samples the combined activity of BMP4, BMP6 and BMP9 could not account for the total BMP activity in these samples. We therefore conclude that FCS may also contain at least one other BMP isoform (BMPX, Figure 6). Remarkably, our studies reveal that FCS contains high levels of BMP activity and that typical cell culture conditions of 5–10% FCS contain BMP activity from 0.5–2 ng/ml of BMPs. Given the profound biological effects of BMP signalling on cell biology and the interaction of these events with other signalling cascades, we believe that it is important to consider the potential effects of BMP signalling when assessing tissue culture experimental results.
As well as measuring BMPs in FCS we expanded our analysis and found that BMP9 is present in human serum in 1–2 ng/ml quantities in concordance with recently published findings . To the best of our knowledge we identify BMP4 as a human serum factor for the first time. We failed to detect BMP6 or any other BMP activity capable of inducing BRE-Luc in human serum indicating that BMP4 and BMP9 maybe the only circulating BMPs present in normal human serum.
An increasing body of evidence is revealing that aberrant BMP signalling and changes in BMP levels may have important biological consequences during normal and pathophysiological processes . We have demonstrated that our bioassay is capable of measuring the activity of individual BMP isoforms from serum samples. It is an exciting possibility that determination of BMP levels in patient biological samples such as serum, or body cavity fluids may have prognostic and/or diagnostic utility in the management of human disease. Recent studies have also suggested the possibility of using recombinant BMPs to treat osteogenic disease . The use of BMP bioassays may also be useful in monitoring the bioactivity of BMPs in these clinical settings. Our assay could also be used in determining the specificity and efficacy of agents designed to target BMP signalling. With the increasing availability of commercial BMP blocking antibodies and recombinant proteins it should be possible to broaden the applicability of our bioassay. Furthermore, by extrapolating the methodology we describe new cell lines capable of measuring further isoforms of BMP for the use in laboratory research and clinical practice could be developed.