Research Article: Identification of CREB3L1 as a Biomarker Predicting Doxorubicin Treatment Outcome

Date Published: June 25, 2015

Publisher: Public Library of Science

Author(s): Bray Denard, Andrea Pavia-Jimenez, Weina Chen, Noelle S. Williams, Harris Naina, Robert Collins, James Brugarolas, Jin Ye, Irina U Agoulnik.


Doxorubicin has been shown to inhibit proliferation of cancer cells through proteolytic activation of CREB3L1 (cAMP response element binding protein 3-like 1), a transcription factor synthesized as a membrane-bound precursor. Upon doxorubicin treatment, CREB3L1 is cleaved so that the N-terminal domain of the protein can reach the nucleus where it activates transcription of genes that inhibit cell proliferation. These results suggest that the level of CREB3L1 in cancer cells may determine their sensitivity to doxorubicin.

Mice transplanted with 6 lines of renal cell carcinoma (RCC) were injected with doxorubicin to observe the effect of the chemotherapy on tumor growth. Immunohistochemistry and bioinformatics analyses were performed to compare CREB3L1 levels in types of cancer known to respond to doxorubicin versus those resistant to doxorubicin.

Higher levels of CREB3L1 protein are correlated with increased doxorubicin sensitivity of xenograft RCC tumors (p = 0.017 by Pearson analysis). From patient tumor biopsies we analyzed, CREB3L1 was expressed in 19% of RCC, which is generally resistant to doxorubicin, but in 70% of diffuse large B-cell lymphoma that is sensitive to doxorubicin. Doxorubicin is used as the standard treatment for cancers that express the highest levels of CREB3L1 such as osteosarcoma and malignant fibrous histiocytoma but is not generally used to treat those that express the lowest levels of CREB3L1 such as RCC.

Identification of CREB3L1 as the biomarker for doxorubicin sensitivity may markedly improve the doxorubicin response rate by applying doxorubicin only to patients with cancers expressing CREB3L1.

Partial Text

Personalized medicine has gained much attention for cancer treatment and has set off a search for biomarkers that identify individuals who will benefit from a particular form of treatment. Doxorubicin has been extensively used to treat various cancers, but the response rate of the treatment for most cancers is low owing to the lack of such a biomarker to identify patients who will likely to benefit from the treatment. Doxorubicin has been assumed to exert its cytostatic effect through DNA damage, but a consistent correlation between doxorubicin-induced DNA breaks and secession of cell proliferation has not been established [1,2].

In order to study the effect of doxorubicin on growth of xenograft tumors transplanted in mice, we first determined the dose of doxorubicin that can be safely applied to mice and compared it to that usually applied to human patients. Important pharmacokinetic parameters such as Cmax and AUC of the mice receiving a single injection of doxorubicin at 5 mg/kg were lower than that of human patients receiving a single injection of doxorubicin at a typical clinical dose [7] (Table 1). However, doxorubicin appeared to be more toxic to mice than humans, as this dose of doxorubicin killed mice within a week after the injection. We then lowered the dose, and found that mice can tolerate up to 8 injections of doxorubicin at 0.75 mg/kg. Pharmacokinetic modeling analysis indicated that the cumulative therapeutic concentration of the drug under this condition was significantly lower than that in human patients treated with doxorubicin (Table 1, AUC).

The current study demonstrates that production of CREB3L1 in tumor cells is crucial to determine their sensitivity to doxorubicin. Using mouse xenograft models, we showed that doxorubicin shrank tumors expressing high levels of CREB3L1 but not those that had low or no expression of the protein. This conclusion was further supported by our observation that the percentage of the tumors expressing CREB3L1 is higher in cancer types empirically determined to be responsive to doxorubicin treatment (e.g. DLBLC) as opposed to those resistant to doxorubicin (e.g. RCC). When we analyzed DLBLC, we also tried to determine whether higher expression of CREB3L1 in tumor biopsies before chemotherapy was correlated to better responses to later treatment with doxorubicin. The result was inconclusive because these cancer patients were not treated with doxorubicin alone but with chemotherapy regime containing multiple drugs.