Monthly Archives: February 2013

More is not always merrier- methylation version.

Hi there,

This is a mini paper-review of sorts. A few posts ago I described work by Robert Gatenby’s work showing that using lower doses of chemotherapy could potentially lead to more durable antitumour responses. Today I will review some rather intriguing findings using a drug called decitabine, again starting off with the premise that lower doses might be better than higher doses.

Drugs like Azacitidine and Decitabine are passive hypomethylating agents – they are analogs of cytosine – one of the four bases that make up DNA, and react with DNA methyltransferase 1, which carries out DNA methylation and traps it. With no free enzyme to copy methylation states to daughter cells, methylation is progressively lost, with the hope being that genes that have to be silenced by methylation for cancer cells to survive/stay undifferentiated are reactivated, resulting in the castration of a cell’s malignant potential.

These drugs have found extremely limited use due to very high toxicity – being limited to haematological conditions like Myelodysplastic syndrome, but people noticed that there were delayed effects that appeared to be independent of the cytotoxicity of high doses. This, combined with a lung cancer trial where low doses in combination with an HDAC inhibitor showed some good results led researchers to investigate what treating cells with extremely low, normally subtoxic doses of these agents could do.

They found that the agents at those doses could blunt the ability of treated leukamia cell lines to form colonies in glass dishes (that’s a standard assay we use to test how effective an anticancer drug might be) and to induce leukaemia when injected into NOD/SCID mice (which lack an immune system and will therefore accept grafts of human cancers). Importantly, these doses did not hit the colony forming ability of normal blood-forming cells.

low-dose decitabine inhibits colony formation in leukaemia cell lines (a), reduces engraftment of leukaemia cells in bone marrow in NOD/SCID mice (b) and reduces the number of CD34+ cells that have engrafted (CD34 is a surface protein that marks the leukaemic stem cell compartment) (c)

low-dose decitabine inhibits colony formation in leukaemia cell lines (a), reduces engraftment of leukaemia cells in bone marrow in NOD/SCID mice (b) and reduces the number of CD34+ cells that have engrafted (CD34 is a surface protein that marks the leukaemic stem cell compartment) (c)

They did similar experiments with 5-Aza and Decitabine using a breast cancer cell line (MCF-7) and primary breast cancer tissue xenografts. Whatever was true for leukaemia cell lines was also true for these samples. They did a range of primary and secondary xenografts, wherein cells were treated with the drug for 72 hours at low doses (replaced daily) and left for 7-14 days before being put into mice, and secondary transplantations were made from these. There were potent long term effects brought about by drug treatment compared to untreated cells.

Decitabine treatment results in potent long term effects on tumour growth in both primary and secondary xenografts relative to untreated cells, suggesting effects through epigenetic reprogramming.

Decitabine treatment results in potent long term effects on tumour growth in both primary and secondary xenografts relative to untreated cells, suggesting effects through epigenetic reprogramming.

In primary breast cancer samples, they found that colony formation was impaired and treatment with Azacitidine depleted cells in the stem-like cell compartment (CD44+, CD24-). They profiled methylation and expression changes using arrays in treated cell lines and found a host of changes associated with increased expression of genes that suppress the cell cycle proper and entry into it.

I’ve tried using the drug on some of the cell lines I’ve worked on in the past and saw similar results in how treatment with a very low dose of decitabine could, upon transient exposure, result in cells that show divergent growth properties. Clearly, decitabine might be a promising addition to therapy in a broad spectrum of tumour types, especially if there are driver methylation events that drive cell survival. The existence of such methylation states will be the subject of another blog post soon in the future.

Paper Reference –
Tsai HC, Li H, Van Neste L, Cai Y, Robert C, Rassool FV, Shin JJ, Harbom KM,
Beaty R, Pappou E, Harris J, Yen RW, Ahuja N, Brock MV, Stearns V, Feller-Kopman D, Yarmus LB, Lin YC, Welm AL, Issa JP, Minn I, Matsui W, Jang YY, Sharkis SJ,Baylin SB, Zahnow CA. Transient low doses of DNA-demethylating agents exert durable antitumor effects on hematological and epithelial tumor cells. Cancer Cell. 2012 Mar 20;21(3):430-46. doi: 10.1016/j.ccr.2011.12.029. PubMed PMID:22439938; PubMed Central PMCID: PMC3312044.

Cheers,
Ankur ‘Exploreable’ Chakravarthy.

Advertisements

Paper Review: Identification and functional validation of HPV-mediated hypermethylation in head and neck squamous cell carcinoma

Hello everyone.

It has been bloody long since I last blogged because I have been battling illness and getting stuck into research at the same time. I’m going to review a paper I did some work towards in this post. I have written about DNA methylation in the past and the research groups I worked with/continue to work with were focusing on the various genetic and epigenetic alterations that characterise  head and neck squamous cell carcinoma (HNSCC), which is the sixth most common type of cancer.

HNSCC can be thought of as two distinct cancers with vastly different prognosis and aetiologies; a vast majority are caused by heavy smoking overlapping with heavy drinking and an increasing proportion is caused by HPV infection, transmissible through oral sex. (HPV, by the way, is the same virus that drives cervical cancer). HPV infection per se is insufficient to cause the cancers associated with it – there have to be additional genetic and epigenetic modifications on top. HPV positive HNSCC has excellent survival relative to HPV negative HNSCC, by the way.

In this study, we obtained clinical samples of both HPV positive and HPV negative HNSCC – some were fresh frozen upon surgical resection/biopsy, a lot were FFPE samples (Formalin fixed,paraffin embedded) and we also profiled cell lines using Illumina 450k methylation arrays, which give a read-out of methylation at 483,000 CpG sites (A cytosine followed by a guanine) across the human genome for less than a large pizza per sample.

The FFPE samples were used as a training set and the Fresh frozen samples and cell lines were used as a validation set. We found quite a few interesting things…

[1] HPV positive HNSCC exhibits a greater degree of DNA methylation (Hypermethylation) than HPV negative HNSCC, especially in genes that are known to be silenced by PRC2 complexes in stem cells. PRC2 complexes consist of multiple proteins that co-operate to produce the H3k27 histone mark. We find the same genes being silenced by DNA methylation instead in HPV positive HNSCC. This is also strongly associated with differences in expression; More the methylation, less the expression, as it should be.

HPV positive HNSCC exhibits hypermethylation relative to HPV negative HNSCC. Blue represents high methylation and yellow represents low methylation.

HPV positive HNSCC exhibits hypermethylation relative to HPV negative HNSCC. Blue represents high methylation and yellow represents low methylation.

[2] A subset of HPV positive HNSCC showed very high degrees of methylation – which is called a CpG Island Methylator Phenotype, and was associated with significantly worse survival.

CIMP phenotype (Cluster 1a) is associated with very high methylation, HPV positivity and significantly worse survival compared to HPV positive tumours with comparatively less methylation (Cluster 1b) as shown in the Kaplan Meier curve at the right.

CIMP phenotype (Cluster 1a) is associated with very high methylation, HPV positivity and significantly worse survival compared to HPV positive tumours with comparatively less methylation (Cluster 1b) as shown in the Kaplan Meier curve at the right.

[3] If you put the viral oncogenes E6 and E7 into a cell line that was derived from HPV negative HNSCC, you tend to see that E6 induces hypermethylation. This wouldn’t be surprising because p53, which is blocked by E6, is known to regulate DNMT1, a DNA methyltransferase that is involved in the maintenance of methylation.

[4] If you use probes on the array that are significantly different between HPV positive HNSCC and HPV negative HNSCC, and compare them to publicly available data for cervical and lung cancer by a process called multidimensional scaling, you find that HPV negative HNSCC is closely related to lung cancer while HPV positive HNSCC is closer to cervical cancer, suggesting that HPV modulates the methylation patterns that make cervical cancer closer to HNSCC of this type.

Multidimensional scaling shows HPV negative HNSCC (HPV0) to be more closely related to lung cancer and HPV positive HNSCC to be similar to cervical cancer.

Multidimensional scaling shows HPV negative HNSCC (HPV0) to be more closely related to lung cancer and HPV positive HNSCC to be similar to cervical cancer.

[5] The relationship between methylation and expression is valid and as predicted even in a panel of HNSCC cell lines, as I demonstrated using qPCR, where we get RNA, make DNA, and then do PCR to find out how many cycles it takes to get past a particular threshold of fluorescence.

Genes that are significantly hypermethylated in HPV positive HNSCC are relatively overexpressed in HPV negative HNSCC as expected (The relationship between most methylation and expression is inverse)

Genes that are significantly hypermethylated in HPV positive HNSCC are relatively overexpressed in HPV negative HNSCC as expected (The relationship between most methylation and expression is inverse)

[6] We found that DNMT1 and DNMT3a , which are enzymes involved in maintaining and establishing DNA methylation, are expressed more in HPV positive HNSCC cell lines relative to HPV negative HNSCC cell lines as a group.

DNMT1 and DNMT3a are significantly overexpressed in a panel of HPV positive HNSCC cell lines vs HPV negative HNSCC cell lines.

DNMT1 and DNMT3a are significantly overexpressed in a panel of HPV positive HNSCC cell lines vs HPV negative HNSCC cell lines.

So basically, we started off with two subsets of a type of cancer, identified that the methylation patterns between them are different, that this has functional ramifications and clinical implications. It would be very interesting if someone ended up looking at hitting methylation in HPV positive cancers with anti-methylation drugs to see if that high level of methylation is just an artefact of how HPV positive HNSCC develops or whether there are therapeutic opportunities to be had.

Journal Reference 
Matthias Lechner, Tim Fenton, James West, Gareth Wilson, Andrew Feber, Stephen Henderson, Christina Thirlwell,Harpreet K Dibra, Amrita Jay, Lee Butcher, Ankur R Chakravarthy, Fiona Gratrix, Nirali Patel, Francis Vaz, Paul O’Flynn, Nicholas Kalavrezos, Andrew E Teschendorff, Chris Boshoff and Stephan Beck, Identification and functional validation of HPV-mediated hypermethylation in head and neck squamous cell carcinoma, Genome Medicine 2013, 5:15 doi:10.1186/gm419
URL – http://genomemedicine.com/content/5/2/15

Cheers,
Exploreable.