A trio of papers has provided new hints into how best to use small molecule activators of pyruvate kinase M2 isozyme, a cancer metabolism enzyme present nearly universally across tumor types. Based on that research, activating the enzyme might be most effective in tumors with low levels of serine or oxygen.1-3

Pyruvate kinase occurs as two isoforms: pyruvate kinase M1 isozyme (PKM1) is expressed by most nonproliferative healthy tissues and PKM2 is mainly expressed in cancer cells. Both enzymes catalyze the final step in the energy-producing process of glycolysis.

The major difference between PKM1 and PKM2 is that PKM1 is constitutively active, whereas PKM2 has low basal activity that can be fine-tuned up or down by other metabolic and signaling pathways. Although both enzymes convert phosphoenolpyruvate into ATP and pyruvate, the final product of glycolysis, PKM2 does so at a much slower rate than PKM1.

In cancer cells expressing primarily PKM2, this results in a bottleneck in the glycolytic pathway that leads to the buildup of intermediate metabolites of glycolysis, which are believed to feed alternative biosynthetic pathways that enhance tumor cell growth and proliferation.

Based on these observations, researchers at Harvard Medical School hypothesized that activating PKM2 in cancer could reduce tumor growth. Indeed, a 2008 paper in Nature showed that replacing PKM2 in cancer cells with the more active PKM1 decreased tumor formation in mice.4

The challenge has been identifying small molecule activators of PKM2.

Now, a team from The Beatson Institute for Cancer Research and Astex Pharmaceuticals Inc. has identified a previously undescribed allosteric site on PKM2 that binds serine to boost activity of the enzyme.1 Independently, teams from the Massachusetts Institute of Technology and Agios Pharmaceuticals Inc. have each identified small molecule activators of PKM2 that bind a distinct allosteric site from the serine binding pocket and have shown that these molecules reduce cancer cell proliferation in combination with particular metabolic stresses.2,3

Because activating PKM2 was only effective in blocking cancer cell growth under very specific metabolic conditions, the key next step will be identifying cancer subtypes or combination treatments that elicit a similar metabolic stress to sensitize tumors to the PKM2 activators.

Agios has a preclinical-stage program targeting PKM2 in cancer, and Astex has a discovery-stage program targeting PKM2 in cancer.

The serine link

The Beatson-Astex team first knocked down both pyruvate kinase isoforms in colon cancer cells to see which metabolic biosynthesis pathways were affected. In cancer cells with reduced glycolysis, flux through the serine biosynthetic pathway was increased.

The team next asked if there was pathway feedback from serine biosynthesis back to glycolysis. Growing the cancer cells in serine-depleted media, which causes the cells to upregulate serine biosynthesis, decreased glycolytic activity compared with growing cells in serine-rich media.

Finally the team looked to see whether serine might directly connect the two pathways. Indeed, serine bound to PKM2 in vitro and activated enzyme activity. Crystallographic experiments showed that serine bound at a previously undescribed pocket.

Thus, serine is a direct connection between the two pathways. When serine levels are low, PKM2 activity is low and metabolites are shunted into the serine biosynthesis pathway. When serine levels are high, the metabolite activates PKM2, which increases glycolysis and decreases serine biosynthesis.

The results identify a site that can be targeted to activate PKM2 and also suggest that activating PKM2 when serine levels are low might short-circuit an adaptive tumor response.

Results were reported in Nature.

Activation energy

Teams led by Agios and MIT researchers started with a small molecule activator of PKM2 and looked for metabolic conditions in which the activating molecule lowered cancer cell proliferation.

The Agios team, led by Senior Director of Biochemistry Lenny Dang, ran a high throughput screen and medicinal chemistry optimization to identify a small molecule activator of PKM2.

In a lung cancer cell line, the PKM2 activator reduced production of serine, suggesting that flow through the glycolysis and serine biosynthesis pathways were linked.

Thus, the Agios team looked directly at the impact of amino acid availability on the anticancer activity of its PKM2 activator. The molecule decreased cancer cell proliferation when the cells were grown in media lacking serine but not when the cells were grown in serine-replete media.

An MIT team led by Matthew Vander Heiden also explored the conditions under which boosting PKM2 activity affected cancer cell growth. To do this, the team turned to TEPP-46, a small molecule activator of PKM2 it identified in 2010.5 Vander Heiden is an assistant professor of biology at MIT.

Similar to what Agios found, the effects of the MIT team's PKM2 activator depended on cell growth conditions. TEPP-46 decreased cancer cell proliferation under hypoxic conditions but not under aerobic conditions.

The Agios and MIT molecules both activated PKM2 by binding at the subunit interface-a different site from where serine binds.

Finally, the MIT team asked if the PKM2 activator would reduce tumor growth in vivo. In a xenograft mouse model of lung cancer, TEPP-46 increased tumor latency and decreased tumor size compared with vehicle.

The MIT results were published in Nature Chemical Biology. Agios reported its findings in Chemistry & Biology.

"My take-home from this is that the degree of PKM2 activity can set a cancer cell in an ATP-producing state that is not conducive to growth. When PKM2 activity is high, cells do not engage in growth-promoting metabolism, and this state is tumor suppressive," said Vander Heiden, who sits on Agios' scientific advisory board.

He cautioned that tumor suppression is distinct from shrinking existing tumors.

Vander Heiden added that "in cancer metabolism, we are still trying to understand the critical metabolic nodes that cancer cells depend on. These studies suggest PKM2 may be a key controller in some cases, but we still do not know which tumors will be more sensitive. The emerging connection to serine may be one avenue for figuring out which patients might respond."

"The big challenge with cancer metabolism targets to date is that none have had a sufficiently dramatic effect in preclinical models to be exciting enough to take forward-and that's the case with PKM2. There are effects, but they are small. This is likely because there are homeostatic mechanisms to guard against just this kind of pathway perturbation," said Neil Thompson, SVP of biology at Astex.

Thus, Thompson thinks treating cancer with small molecules that modulate metabolic pathways will likely require either a tumor genetic alteration or combination treatment that overrides these homeostatic mechanisms.

Putting PKM2 into context

Vander Heiden's group is now testing the effects of PKM2 activators in genetically engineered mouse cancer models and asking if the molecules are effective against existing tumors and if there is differential sensitivity across different tumor models.

Eyal Gottlieb, who was co-lead of the Nature study, suggested that combining PKM2 activators with serine uptake blockers might be the way to go. Gottlieb is a group leader in apoptosis and tumor metabolism at Beatson.

"The good news is that you don't need to deplete serine completely; it is enough to partially deplete it, and there is one high-affinity serine receptor that could be targeted by a small molecule or an antibody," said Gottlieb.

Thompson told SciBX that "ideally, a genetic selection strategy" would be the way to go for patient stratification, adding that "the best-characterized genetic alterations that confer serine deficiency are in breast cancer."

Alternatively, he said a combination strategy that compromises serine biosynthesis might be an option.

Astex has found activators that bind at several distinct sites, but Thompson said the serine site is particularly interesting.

The company has also identified inhibitors of PKM2. "In our thinking we've always felt that inhibitors could be as good as activators but would require targeting tumors with a different metabolic footprint-tumors that have a high dependence on the glycolytic pathway," he added.

TEPP-46 was discovered under a collaboration between MIT and the NIH. The NIH holds issued composition-of-matter patents covering this molecule and an additional PKM2 activator series. The IP is available for licensing.

Agios declined to discuss its PKM2 program, but Dang did say that "PKM2 is still important in our cancer metabolism research portfolio."

Agios' lead programs target isocitrate dehydrogenase 1 (IDH1) and IDH2, which are mutated in various cancers and lead to increased production of the oncogenic metabolite 2-hydroxyglutarate

Dynamix Pharmaceuticals Ltd. has a small molecule PKM2 activator, DNX-3000, in preclinical development for cancer.

Kotz, J. SciBX 5(43); doi:10.1038/scibx.2012.1127
Published online Nov. 1, 2012


1.   Chaneton, B. et al. Nature; published online Oct. 14, 2012; doi:10.1038/nature11540
Contact: Eyal Gottlieb, Cancer Research UK and The Beatson Institute for Cancer Research, Glasgow, U.K.
e-mail: e.gottlieb@beatson.gla.ac.uk
Contact: Marc O'Reilly, Astex Pharmaceuticals Inc., Cambridge, U.K.
e-mail: marc.oreilly@astx.com

2.   Kung, C. et al. Chem. Biol.; published online Sept. 21, 2012; doi:10.1016/j.chembiol.2012.07.021
Contact: Lenny Dang, Agios Pharmaceuticals Inc., Cambridge, Mass.
e-mail: lenny.dang@agios.com

3.   Anastasiou, D. et al. Nat. Chem. Biol.; published online Aug. 26, 2012; doi:10.1038/nchembio.1060
Contact: Matthew G. Vander Heiden, Massachusetts Institute of Technology, Cambridge, Mass.
e-mail: mvh@mit.edu

4.   Christofk, H.R. et al. Nature 452, 230-233 (2008)

5.   Boxer, M.B. et al. J. Med. Chem. 53, 1048-1055 (2010)


Agios Pharmaceuticals Inc., Cambridge, Mass.

Astex Pharmaceuticals Inc., Dublin, Calif.

The Beatson Institute for Cancer Research, Glasgow, U.K.

Dynamix Pharmaceuticals Ltd., Rehovot, Israel

Harvard Medical School, Boston, Mass.

Massachusetts Institute of Technology, Cambridge, Mass.

National Institutes of Health, Bethesda, Md.