An imbalance between osteoblasts and osteoclasts results in osteolytic bone disease, a major complication in multiple myeloma (MM). Using gene expression profiling (GEP), we identified CYR61 as the most significantly overexpressed gene in bone biopsies in subsets of MGUS and MM patients compared with healthy controls. CYR61 is induced during osteoblast differentiation by endothelin-1, (Clines et al., 2007) and by Wnt3a (Si et al., 2006). CYR61 also inhibits the formation of multinucleate osteoclasts in vitro (Crockett et al., 2007). GEP revealed that elevated levels of CYR61 in 90/264 (34%) of newly diagnosed total therapy 3 (TT3) patients is associated with superior overall survival (p=0.006, HR-0.78). The CYR61 protein, CCN1, is significantly elevated in 57/177 (32%) TT3 patients at diagnosis and is associated with superior overall survival (p=0.024, HR-0.1). Recombinant CCN1 reduces MM cell growth and survival in vitro, and ectopic CYR61 produced by MM cells implanted into the SCID-hu model show that CCN1 inhibits osteolysis and reduces MM tumor growth in vivo. Taken together these data suggest that elevated CCN1 may be indicative of a dysfunctional microenvironment (ME) and could serve as a biomarker to monitor the MM ME.
To test whether CCN1 could serve as an MM ME biomarker, CCN1 protein was quantitated by ELISA (DRG International). In BM sera from patients with MGUS/AMM (n=84), newly diagnosed myeloma (n=239), and in paired samples at diagnosis and 48h after a test dose of velcade (n=168 pairs), in nCR/CR (n=156 pairs) or at relapse (n=41 pairs). Differences between groups were analyzed using the t-test and survival using the log-rank test.
CCN1 levels in myeloma patients at diagnosis (464.5 ± 30.4, p=0.0007), near complete remission (nCR, 466.1 ± 51.9 pg/ml, p=0.0055), complete remission (CR, 510.4 ± 36.5 pg/ml, p< 0.0001) and at relapse (692.5 ± 166.2 pg/ml, p=0.0185) were significantly elevated compared with healthy donors (274.4 ± 40.1 pg/ml). In paired samples, CCN1 levels at 48h post test-dose of velcade were significantly higher than at diagnosis (p=0.005). This suggests that elevated levels of CCN1 post-velcade may reflect recovery of the MM ME. To determine whether CCN1 may be associated with CR duration, CCN1 levels at nCR/CR were compared in patients that remained in CR for < 5 years and ≥ 5 years. CCN1 levels were significantly elevated in patients in CR ≥ 5 years (563.2 ± 53.2 pg/ml, p=0.01) compared to those in CR < 5 years (411.4 ± 22.8). CCN1 levels at nCR/CR and molecularly-defined risk were examined to determine whether they were independently associated with survival. Four groups were compared: CCN1 quartiles 1–3 levels [Q1-Q3] with low- or high-risk by GEP 70 or CCN1 [Q4] with low- or high-risk. While CCN1 levels were not prognostic for low-risk patients, Q4 high-risk patients had significantly longer PFS (Figure 1a, p=0.0120) and OS (Figure 1b, p<0.0001) than those with Q1-Q3 levels (median OS 6.9 and 3.8 years, respectively).
To determine whether baseline CCN1 levels are associated with time to CR, CCN1 levels in patients who had not achieved CR (368.0 ± 38.0 pg/ml n=25) were compared to those that achieved CR within 6 months (525.6 ± 60.0 pg/ml, p=0.0284) and within 1 year (548.2 ± 74.5 pg/ml n=96, p=0.033). Patients who achieve CR had significantly higher CCN1 levels than those who did not reach CR. At diagnosis, MM CCN1 levels were lower than those of MGUS/AMM (608.3 ± 71.5 pg/ml, p=0.0324). We thus examined CCN1 levels in untreated MGUS and AMM patients to determine whether they were associated with time to progression to overt MM. Patients who progressed in less than 3 years had CCN1 levels significantly lower (441.6 ± 28.5 n=38) than those who progressed later (726.7 ± 121.7 pg/ml n=46, p=0.0399). CCN1 levels at the upper quartile were associated with significantly longer time to progress to overt MM (p<0.0056).
Taken together, these data suggest that elevated CCN1 levels indicate a recovery effort by the ME and due to its positive association with survival of myeloma patients may be useful in the monitoring of disease course and thus personalized therapeutic management.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.