CDK1/2 toolbox in need of an upgrade

This article refers to:

Critical reanalysis of the methods that discriminate the activity of CDK2 from CDK1

Cyclin-dependent protein kinases are key drivers of cell cycle progression.^1,2 Specific phases of the cycle are controlled by altering the levels and availability of cyclins which function to bind and activate their CDK partners. The timely synthesis and rapid destruction of individual types of cyclin ensures an orderly and unidirectional advance through the cell cycle. Because the majority of cancers are characterized by alterations that deregulate growth control pathways, particularly the R-point governing the G1 to S transition, CDKs are important targets for cancer therapy.^2,3

As with any research area, and especially when a family of highly conserved proteins is the focus, data quality is critically dependent on the quality of the available reagents. In a recent Cell Cycle article,⁴ the specificity of certain reagents used to distinguish CDK1 and CDK2 has been questioned. These problems surfaced when the authors were studying the sensitivity of different cancer cell lines toward the Chk1 inhibitor, MK-8776.⁵ A major question raised is the specificity of commercial antibodies directed against phosphotyrosine 15 (PY15) on CDK1 and CDK2. This is an inhibitory site of phosphorylation that requires reversal by CDC25 phosphatases for CDK activation.¹ Sakurikar and Eastman⁴ compile a list of 27 such antibodies from 12 companies, and in only 5 cases do the data-sheets indicate cross-reactivity. This leaves a startling 22 antibodies purported to be selective either for PY15-CDK1 or PY15-CDK2 that, given the high degree of sequence conservation in this region of CDKs,⁴ are almost certainly not selective. To illustrate this point, CDK2 was immunoprecipitated from U2OS cells, and found to react not only with an antibody to CDK2, as expected, but also with a purported phospho-specific antibody to PY15-CDK1, and this despite an absence of CDK1 in the immunoprecipitate.⁵ While the reciprocal experiment, to immunoprecipitate CDK1 and determine if it reacts with an antibody against PY15-CDK2, was not presented, the cross-reactivity of the PY15-directed antibodies highlighted in the analysis should encourage caution in their use.

The development of selective CDK inhibitors has been of immense value for advancing our knowledge of CDK function, and in the development of novel cancer therapeutics.^1-3 Sakurikar and Eastman⁴ make a case that CDK1 and CDK2 inhibitors are rarely selective. Two main points are made. The first is that many purportedly specific CDK inhibitors are only selective in a narrow concentration range based on in vitro assays, and often inhibit other types of CDKs at marginally higher concentrations. One example cited is dinaciclib, which is marketed as a CDK2 inhibitor, yet inhibits several other CDKs with similar potencies.^1,3 The second point addresses whether selectivity in vitro can be extrapolated to cells. The examples provided in the analysis indicate that the answer to this question can be yes or no. For instance, Ro3306, reported to be 10-fold more selective for CDK1 than CDK2 in vitro ⁴, was found to inhibit both kinases equally in cells. On the other hand, CVT-313, reported to be 10-fold more selective for CDK2 than CDK1⁴, did indeed show the expected specificity in cells, inhibiting CDK2-mediated phosphorylation events at concentrations ineffective for the inhibition of CDK1. While reagents more specific for CDK1 or CDK2 would be of obvious value, the authors were able to establish a role for CDK2 rather than CDK1 in their system⁵, through the use of multiple existing reagents, and perhaps more importantly, by an awareness of the limitations of those reagents.

A final point in the commentary⁴, concerning assessment of CDK2 activity in cells, is that elevated cyclin E levels are often equated with CDK2/cyclin E being active, e.g. Ref.⁶ However, since CDK2/cyclin E in partnership with GSK3β phosphorylates cyclin E marking it for degradation⁷, high cyclin E levels may in fact reflect low CDK2/cyclin E activity.^4,5 While possible solutions to this dilemma were not discussed, it seems likely that conclusions drawn about CDK2/cyclin E activity based solely on assessment of cyclin E levels may be flawed.

The article underscores several concerns in the methods to discriminate CDK1 from CDK2, and provides a basis for reevaluation of existing data and more rigorous interpretations of future results.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.