Harvard scientists reveal how cortistatin A kills leukaemia cells

Cortistatin A (CA) is a natural steroidal alkaloid isolated from sea sponges[1]. Kyriacos Nicolaou totally synthesized the compound and characterized it as an inhibitor of CDK8 and CDK19[2]. Yesterday, a group of Harvard scientists led by Matthew Shair published a paper in Nature[3], revealing the specific mechanism of CA killing leukaemia cells.

It is well known that super-enhancers (SEs) play an important role in tumor pathogenesis. The BRD4 inhibitor I-BET151 downregulated SE-associated genes, thereby inhibiting cancer cells. However, the new paper suggested that CA upregulated these SE-associated genes, yet also inhibited leukaemia cells. Shair believes that cancer cells are sensitive to the dosage of SE-associated genes.

The authors tested CA in approximately 400 kinases, and found that the compound inhibits only CDK8 (IC50=12 nM) and CDK19 in cells. The crystal structure of CA/CDK8 complex (Protein Data Bank ID: 4CRL) reveals the exquisite binding mode. Most kinase inhibitors do not have this type of 3D structure.

Cortistatin A

The researchers measured the efficacy of CA in an AML model using SET-2 cells. The drug afforded a 71% tumour volume reduction, also with no loss in body weight.

CA was totally synthesized via 32 chemical steps. It should be very difficult to make the compound on a large scale. Scientists are trying to find less complex structures that work just like the natural compound.

[1] J Am Chem Soc. 2006, 128(10), 3148-3149.
[2] Angew Chem Int Ed Engl. 2009, 48(47), 8952-8957.
[3] Nature. 2015, doi: 10.1038/nature14904.


Adding pioglitazone to imatinib may help fight leukemia

Gleevec (imatinib) is a breakthrough therapy that turns deadly chronic myeloid leukemia (CML) into a manageable disease. However, only 5% of patients achieved complete molecular response (CMR)[1] that no Bcr-Abl mRNA is detectable in the marrow.

A group of French researchers added Actos (pioglitazone) to imatinib on two patients diagnosed with both diabetes and CML[2]. Both two patients had never reached CMR in spite of long-term imatinib treatment. Amazingly, they achieved CMR after 1 year of pioglitazone addition.

pioglitazone plus imatinib

Before filing a formal clinical trial application, the researchers prescribed pioglitazone off-label to a third CML patient, this time non-diabetic, who never reached CMR either under long-term imatinib therapy. Finally, CMR was achieved after 6 months of pioglitazone addition.

All patients achieved sustained CMR up to 4.7 years after withdrawal of pioglitazone.

The authors note that activation of PPARγ by pioglitazone reduces expression of STAT5, while imatinib inhibits STAT5 activation by Bcr-Abl pathway. Imatinib alone is sufficient to kill the bulk of cancer cells, but fails to bring STAT5 activity below a threshold for killing cancer stem cells. Pioglitazone is effective at doing so in synergy with imatinib.

The trial is too small. It is difficult to know if the results would reproduce in a lager trial. It is also unknown whether CMR offers significant survival benefit. Imatinib already enable 90% of patients to survive for 5 years. A long-term, larger trial directly comparing the combination therapy versus imatinib alone is warranted.

[1] J Clin Oncol. 2008, 26(20), 3358-3363.
[2] Nature. 2015, doi:10.1038/nature15248.

CAR-T treated leukemia patients survive for 4.5 years

In the summer of 2010, Novartis and the University of Pennsylvania began testing their anti-CD19 CAR-T therapy, CTL019, in patients with relapsed and refractory chronic lymphocytic leukemia (CLL). The mature results from the trial were published in Science Translational Medicine[1]. Dr. Carl June led the study.

The trial enrolled 14 heavily pretreated CLL patients. These patients were treated with CTL019 at doses of 1.4 × 10^7 to 11 × 10^8 cells. Four patients had a complete response (CR), and additional four patients had a partial response (PR). Six patients did not respond to the therapy and progressed soon.

One of the CR patients died at 21 months due to infections after removal of the skin cancer on his leg. The three other patients are still alive with no signs of recurrence. At the time of the report, they have survived for 28, 52, and 53 months after receiving CTL019.

Two of the PR patients died of disease progression at 10 and 27 months. One of the PR patients died from a pulmonary embolism (肺动脉栓塞) at 6 months. The other patient was switched to other therapies due to the disease progressed at 13 months.

The CR rate in CLL was much lower than that in ALL (acute lymphocytic leukemia). No CLL patient with CR has relapsed, while in a previous ALL trial, 37% of CR patients have subsequently relapsed[2]. The CAR-T cells remain in CLL patients’ blood and retain their ability to hunt for cancer cells beyond four years.

[1] Sci Transl Med. 2015, 7(303), 303ra139.
[2] ASH2014, https://ash.confex.com/ash/2014/webprogram/Paper69932.html

CAR-T with low-affinity shows promising results in mice with solid tumor

Anti-CD19 CAR-T therapy has demonstrated potent clinical efficacy in patients with B-cell leukemia and lymphoma. However, this therapy faces many challenges in the context of solid tumors. Most proteins overexpressed on cancer cells may also be expressed on normal cells. CAR-T cells cannot distinguish cancer cells from normal cells. Five years ago, one patient died after the treatment with anti-HER2 CAR-T due to the expression of HER2 in lung tissues[1].

CD19 is not known to be expressed on any healthy tissue other than B cells. Even though such CAR-T cells attack cancer cells and normal B cells, patients can live without B cell for a long time. Further, CD19 is not expressed on hematopoietic stem cells, and therefore B cells should return when the CAR-T cell is no longer present. This may not be the case with solid tumors.

To make CAR-T therapy applicable to solid tumors, Dr. Laurence Cooper of ZIOPHARM Oncology (NASDAQ: ZIOP) developed CAR-T with reduced affinity, and showed that these CAR-T cells could distinguish cancer from normal cells[2]. In other words, Cooper’s CAR-T cells could minimize the “on target off tumor” toxicity in mice.

The researchers chose wild-type EGFR as a target. The protein is overexpressed glioblastoma but is also found at low levels on certain normal cells. Two CARs were generated from cetuximab (cetux) and nimotuzumab (nimo), respectively. Cetux has higher affinity for EGFR, while nimo has low lower affinity.

The researchers then tested the anti-EGFR CAR-T cells in NSG mice with glioma. Cetux-CAR-T cells and nimo-CAR-T cells both significantly inhibited tumor growth. However, the cetux-CAR-T cells caused significant toxicity, resulting in significant death of mice within 7 days of T-cell infusion. The nimo-CAR-T cells have no apparent toxicity.

EGFR CAR T cells

NSG mice lack mature T cells, B cells, and NK cells, and are also deficient in multiple cytokine signaling pathways. It is unclear whether CAR-T with low-affinity could reduce cytokine storm, a common life threatening side effect associated with CAR-T therapy. The researchers used cancer cells that express a median density of 240,000 molecules of EGFR/cell to establish mouse model. In patients, it should be far more complicated.

[1] Mol Ther. 2010, 18(4), 843-851.
[2] Cancer Res. 2015, doi: 10.1158/0008-5472.CAN-15-0139.

Novel blood test could detect pancreatic cancer early

Pancreatic cancer is the most deadly of any common cancer, because the disease is often diagnosed in advanced stages and only 15% of patients qualify for surgery. Scientists from MD Anderson Cancer Center found a novel blood test, which could diagnose pancreatic cancer early[1].

Cells often secrete some small particles containing proteins and nucleic acids into the blood. These particles are called exosomes. The researchers found that exosomes from cancer cells harbored high levels of glypican-1 (GPC1), a cell surface proteoglycan.

The researchers analyzed blood samples from 190 pancreatic cancer patients, 32 breast cancer patients and 100 healthy donors. All pancreatic cancer patients revealed high levels of GPC1 positive circulating exosomes (GPC1+crExos).


BPD: benign pancreatic disease; PCPL, pancreatic cancer precursor lesions; PDAC, pancreatic ductal adenocarcinoma.

GPC1+crExos are superior to the commonly used CA19-9 biomarker. Patients with benign pancreatic disease (BPD) exhibited similar GPC1+crExos levels to healthy donor. However, CA19-9 levels were significantly increased in patients with BPD.

Furthermore, the GPC1+crExos test could detect pancreatic cancer at early stage before the disease could be diagnosed by magnetic resonance imaging (MRI). If pancreatic cancer is diagnosed at an earlier stage, curative surgery could reduce the death rates significantly.

Routine screening for pancreatic cancer using MRI or CT would be prohibitively expensive and associated with a high false-positive rate. The GPC1+crExos test would be feasible and enhance the accuracy of MRI or CT scans.

[1] Nature. 2015, doi:10.1038/nature14581.

A metabolite of abiraterone is more effective than the parent drug

Zytiga (abiraterone) is a blockbuster drug for the treatment of metastatic castration-resistant prostate cancer. The total sales for 2014 reached $2.24 billion. Innocrin Pharmaceuticals and Tokai Pharmaceuticals (NASDAQ: TKAI) are developing me-better versions of abiraterone (VT-464 and galeterone).

Cleveland Clinic researchers found that D4A, a major metabolite of abiraterone, is more effective than the parent drug[1]. D4A inhibits not only CYP17A1, but also 3βHSD and SRD5A, which are required for DHT synthesis. Furthermore, D4A is a competitive, potent androgen receptor antagonist (IC50=7.9 nM).


The researchers examined the effects of D4A in two prostate cancer xenograft models. D4A is more potent than abiraterone for blocking VCaP and C4-2 xenograft progression. Direct treatment with D4A is likely to result in a greater clinical benefit than abiraterone.

[1] Nature. 2015, doi: 10.1038/nature14406.

Merck identifies a new biomarker for anti-PD1 therapy

Several trials have verified that patients whose tumors overexpress PDL1 have improved clinical outcomes with anti-PD1 therapy. At the ASCO2015 Annual Meeting, Merck researchers report a new biomarker called mismatch repair (MMR) deficiency, which predicts superior response to anti-PD1 therapy.

Tumors with MMR deficiency harbor many more mutations than tumors without such repair defects. Whole exome sequencing revealed an average of 1,782 somatic mutations per tumor in MMR-deficient compared to 73 in MMR-proficient tumors. Tumors with more mutations are more likely to be recognized as foreign by the immune system.

In a Phase II trial of pembrolizumab, ORR was 62% in MMR-deficient patients (n=25) compared with 0% in MMR-proficient patients (n=25). Progression-free survival rates at 20 weeks were 78% and 11%, respectively. Merck plans to launch a larger phase II study to confirm these findings.

Approximately 5% of many tumor types have MMR deficiency. About 20% of non-inherited colorectal cancers are MMR deficient. MMR status could be easily determined using an existing commercially available test.