HDAC inhibitors to the rescue in sonic hedgehog medulloblastoma

See the article by Pak et al. in this issue, pp. 1150–1163.

Medulloblastoma (MB) is the most common primary brain cancer in children and comprises 4 molecular subgroups based on transcriptomal analysis: sonic hedgehog (SHH), wingless, Group 3, and Group 4.¹ SHH MB commonly harbors inactivating mutations in Patched 1 (PTCH1), the receptor for SHH, or activating mutations in Smoothened (SMO), a G-protein coupled receptor that is repressed by PTCH1 in the absence of SHH but is de-repressed to activate GLI transcription factors when SHH binds PTCH1 (Figure 1). The cure rate for SHH MB is age dependent, with a 10-year survival of approximately 75% for infants, 50% for children, and 34% for adults.¹ The serendipitous discovery of cyclopamine as a SMO inhibitor almost 20 years ago excited the pediatric neuro-oncology field about the potential to pharmacologically inhibit the hedgehog pathway as a therapy for SHH MB.² Disappointingly, in 2019, SHH pathway inhibitors are still not FDA approved for children with SHH MB but are approved for basal cell carcinoma, a type of skin cancer that also commonly harbors mutations that activate the SHH pathway. There are several potential explanations for this discrepancy. First, SHH MB is a much rarer disease than basal cell carcinoma, and therefore it is more challenging to enroll patients in trials. Second, in children, SHH MB commonly harbors mutations that activate the SHH pathway downstream of SMO, resulting in primary resistance to SMO inhibitors.³ Third, it was noted early on in patients with SHH MB that resistance to SMO inhibition can occur rapidly.⁴ Fourth, SMO inhibition may result in irreversible growth plate closure.⁵ SMO inhibitors were first tested in a recurrent MB patient population and some antitumor activity was noted.^6,7 Currently there is an ongoing clinical trial for children 3 years or greater (NCT01878617) evaluating SMO inhibition as part of the upfront therapy for SHH MB with radiation and multi-agent chemotherapy and we eagerly await the results.

Fig. 1.

HDAC inhibitors as novel SMO-independent SHH pathway inhibitors.

A cartoon depicting the SHH pathway. When SHH binds PTCH1, PTCH1 can no longer inhibit SMO, allowing the GLI transcription factors to enter the nucleus and transcribe SHH target genes. HDAC inhibitors inhibit the SHH pathway either at the level of GLI (by inhibiting the deacetylation of GLI transcription factors) or at the level of histones (by inhibiting the deacetylation of histones that regulate gene transcription of SHH pathway target genes). Either way, HDAC inhibitors inhibit the SHH pathway downstream of SMO.

With this background in mind, Rosalind Segal and colleagues report the results of their unbiased in vitro drug screen to identify promising molecules to treat SHH MB.⁸ The authors use a murine MB cell line derived from a spontaneous MB arising in PTCH1+/− mice. Of note, this cell line was subsequently noted to harbor a p53 mutation, a marker for a more aggressive model. They screen 960 drugs and identify histone deacetylase (HDAC) inhibitors, inhibitors of a class of enzymes that remove acetyl groups from the lysines of histone and non-histone proteins, as potent inhibitors of the SHH pathway. A subsequent validation screen demonstrated that HDAC inhibitors are also active in SMO-inhibitor resistant lines such as those harboring a constitutively active Gli2 or SUFU deletion. While the identification of HDAC inhibitors as a promising therapy against SHH MB is not novel per se,⁹ the observations that HDAC inhibition is efficacious even in SMO-inhibitor resistant models of SHH MB is exciting and novel. As there are 11 HDAC enzymes that are grouped into 4 classes, the authors proceed to identify which specific HDACs are most responsible for activating the SHH pathway in MB. Analysis of human tumors suggest that class I HDACs (which consist of HDAC 1, 2, 3, and 8) are most upregulated in SHH MB. The authors note that knockdown of HDAC1 and HDAC2 using short hairpin RNAs also inhibits the hedgehog pathway. It is worth noting that HDAC1 and HDAC2 have also been implicated in the deacetylation of Gli proteins.¹⁰ Further studies are required to delineate all the mechanisms by which HDAC1 and HDAC2 regulate the SHH pathway (Figure 1). To extend their in vitro observations in vivo, the authors chose to focus on one particularly potent HDAC inhibitor, quisinostat, a pan-HDAC inhibitor. Interestingly, this particular inhibitor has not been tested in children so far, but entinostat, primarily a class I HDAC inhibitor, is currently being evaluated through the Children’s Oncology Group in children with recurrent tumors, including MB (NCT02780804).

It has been approximately 13 years since the first HDAC inhibitor, vorinostat, a pan-HDAC inhibitor, was approved by the FDA to treat cutaneous T-cell lymphoma. While 4 HDAC inhibitors (vorinostat, romidepsin, panobinostat, and belinostat) are currently FDA approved for liquid cancers, none are currently approved to treat a solid cancer. The observations that genetic knockdown of both HDAC1 and HDAC2 is sufficient to inhibit the hedgehog pathway is important as pan-HDAC inhibitors have thus far suffered from excessive toxicity in clinical trials for solid tumors. It is possible that only isoform-specific HDAC inhibitors will prove efficacious without excessive toxicity to treat solid cancers such as SHH MB. Thus, while the data with quisinostat are compelling in the mice, it remains to be determined whether efficacy will be observed with quisinostat in children with SHH MB. In the event that this drug is observed to be toxic when translated into clinical trials for children with SHH MB, additional in vivo studies with brain penetrant HDAC1- and HDAC2-specific inhibitors in SHH MB animal models should be carried out to help prioritize translation of additional HDAC inhibitors into the clinic.