Heat Shock Protein 90 (Hsp90) Activity Status Is An Independent Prognostic Marker in Chronic Lymphocytic Leukemia (CLL) and Correlates with IgV_H Mutational Status, ZAP-70 Expression and Time to First Treatment

Abstract 2544

The clinical presentation of chronic lymphocytic leukemia (CLL) is variable and includes at least two distinct clinical subtypes: indolent or progressive disease. Several prognostic markers such as the mutational status of the IgV_H genes or ZAP-70 expression identify patients that will have more aggressive clinical courses. We have previously reported that ZAP-70 expression in CLL requires the support and stability provided by the Hsp90 activated complex (Hsp90-AC) and that Hsp90 inhibitors induce apoptosis in CLL cells preferentially in patients with high–risk/rapid progressive disease.

Hsp90 requires the conformation of a multimeric complex with other co-chaperones (Hop, p23) to become active and function as a chaperone for mutated, over-expressed, misfolded or constitutively activated proteins. However, to measure Hsp90-AC activity is challenging and requires protein assays that are semi-quantitative and not always can provide a functional read out. To address this problem we designed a test to measure Hsp90-AC by taking advantage of the higher avidity of this complex to bind Hsp90 inhibitors such as 17-AAG. We hypothesize that Hsp90-AC activity is an independent prognostic marker of disease progression in CLL and to test this we measured the ability of 17-AAG to inhibit Hsp90-AC and promote apoptosis in vitro in a cohort of previously untreated CLL patients. This evaluation was performed in samples collected at the time of diagnosis and the results were correlated with the time from diagnosis to first treatment (t-Tx1) as well as the presence of other known prognostic markers.

Ninety-five previously untreated CLL patients were included in this cohort. In vitro sensitivity/apoptosis to Hsp90-AC inhibition using 17-AAG (1μg/ml) was measured by flow cytometry using DiOC6 and PI after 48 hours of incubation. Only samples with more than 60% viability were included and we selected a cut-point of ≥ 53% induction of apoptosis to determine if a sample was sensitive or not to Hsp90-AC inhibition (This cut-point was selected using a ROC analysis for optimal diagnostic performance).

We found that 37 patients (39%) were sensitive to 17-AAG. These patient had a median time from diagnosis to first treatment (t-Tx1) of 7 years compared to 4.2 years in patients that were resistant to 17-AAG (p=0.04). In addition, 37 patients (39%) had unmutated IgV_H genes with a median t-Tx1 of 8.7 years compared to 4.2 years in patients with mutated IgV_H genes (p=0.02) and 41 patients (43%) had high levels of ZAP-70 expression (>20% by flow cytometry) and this group of patients had a median t-Tx1 of 7 years vs. 4.2 years in ZAP-70 negative patients (p=0.03). We evaluated the association between the degree of Hsp90-AC inhibition (17-AGG sensitivity) and IgV_H mutational status and found a strong correlation between these two variables with shorter t-Tx1 in patients that were IgV_H unmutated and sensitive to Hsp90 inhibitor (p=0.0003). We found a similar strong correlation between ZAP-70 expression and sensitivity to Hsp90-AC inhibitor (p=0.003). In our patient cohort, we saw that IgV_H, ZAP-70 and Hsp90 inhibitor sensitivity are independent prognostic markers in CLL (Cox regression, p=0.002). We found that the sensibility and specificity of ZAP-70 test for high-risk CLL (sensitivity 81.1% [95% IC 67–95%] and specificity 77.6%[95% IC 65–90%]) was increased when the Hsp90-AC inhibitor sensitivity test was added (sensitivity 87% [95% IC 71–100%] and specificity 91%[95% IC 79–100%]).

In conclusion, in vitro inhibition of Hsp90-AC in CLL using 17-AAG is an independent prognostic maker and strong predictor of the need for treatment. Hsp90-AC inhibition showed a high correlation with IgV_H mutational status and ZAP-70 expression. Additionally, patients sensitive to 17-AAG whose leukemic cells also highly express ZAP-70 appear to have the worst prognosis with the shortest t-Tx1. Our data suggest that in vitro sensitivity to 17-AAG at the time of diagnosis serves as a quantitative surrogate marker for Hsp90-AC activity and correlates with disease progression in CLL. This is the first time that such correlation has been established in a cohort of cancer patients and shows the relevance of the Hsp90 chaperone system in cancer biology and disease progression.