Background
Cytosolic thymidine kinase 1 (TK1) is a well-known cell cycle-regulated enzyme important for nucleotide metabolism during DNA synthesis [
1]. TK1 catalyzes the conversion of thymidine to deoxythymidine monophosphate, which is further phosphorylated to di- and triphosphates preceding its incorporation into DNA. The activity of TK1 is low or absent in resting cells, increasing in the G
1/S transition and peaking in the S phase, and then disappearing during mitosis [
2‐
5]. Whereas serum TK1 activity is elevated in cancer patients compared with healthy individuals and prognostic in patients with breast and other cancers [
6‐
11], in very few studies have researchers evaluated the utility of serum TK1 for monitoring responses to cancer therapy.
Inhibitors against cyclin-dependent kinase 4/6 (CDK4/6) are an important new class of agents with substantial antitumor activity in patients with advanced hormone receptor-positive (HR+) and human epidermal growth factor receptor 2-negative (HER2−) breast cancer [
12‐
16]. These agents inhibit cell proliferation by activation of retinoblastoma protein, which binds to E2F transcription factors, leading to G
0/G
1 arrest [
17,
18]. The preferential activity of CDK4/6 inhibitors in luminal or HR+ disease is due to the direct link between estrogen receptor (ER) signaling and CDK4/6 activation, because cyclin D is a direct transcription target of ER and other mitogenic signals associated with endocrine resistance [
17,
18].
The potent antiproliferative effect of CDK4/6 inhibition in ER+ breast cancer was demonstrated by tumor Ki-67 analysis in serial biopsies in the NeoPalAna trial (a phase II trial of anastrozole and palbociclib, a CDK4/6 inhibitor, in women with clinical stage II-III ER+/HER2− breast cancer), in which the addition of palbociclib to anastrozole induced complete cell cycle arrest (Ki-67 ≤ 2.7%) in 87% of patients, as compared with 26% following single-agent anastrozole treatment [
19]. Because TK1 is a direct E2F transcription target and is strictly cell cycle-regulated, we hypothesized that changes in levels of TK1 activity before and after administration of a CDK4/6 inhibitor would indicate successful inhibition of CDK4/6, and also that serum TK1 activity could serve as a noninvasive surrogate marker of antitumor activity of CDK4/6 inhibition. The serum samples collected before and after palbociclib in the NeoPalAna trial provide an ideal sample set to test this hypothesis. The objectives of this study were to compare TK1 activity in serum collected before and after anastrozole and palbociclib, and to correlate serum TK1 activity with tumor Ki-67 proliferation index and tumor TK1 mRNA levels in patients with early-stage ER+/HER2− breast cancer enrolled in the NeoPalAna trial.
Discussion
Blood-based biomarkers are of great interest as noninvasive tools for assessing prognosis and in disease monitoring. Circulating tumor cells, exosomes, circulating tumor nucleic acids, and secreted proteins are the most frequently investigated of these potential biomarkers [
24]. However, the clinical utility of such biomarkers is often limited by unsatisfactory sensitivity, specificity, and inter-laboratory reproducibility [
25]. The DiviTum TK1 activity assay has been demonstrated to be a highly sensitive and reliable measurement tool to measure cell proliferation [
6]. Because TK1 is a cell cycle-regulated enzyme that plays a critical role in DNA synthesis, we investigated whether serum TK1 activity could be used as a surrogate marker of the antiproliferative effect of palbociclib in patients with early-stage ER+/HER2− breast cancer enrolled in the NeoPalAna trial (neoadjuvant palbociclib and anastrozole).
This study demonstrates that serum TK1 activity was significantly reduced after 2 weeks of treatment with palbociclib, and that changes in serum TK1 significantly correlated to changes in tumor Ki-67 proliferation index and tumor TK1 mRNA levels. The overall concordance rate in the direction of changes between serum TK and tumor Ki-67 induced by palbociclib was 89.8% (53 of 59, 95% CI 79.2% -96.2%). A reduction in serum TK1 activity by palbociclib had a sensitivity of 94.1% (32 of 34, 95% CI 86.2% -100%) and a specificity of 84% (21 of 25, 95% CI 69.6% -98.4%) in predicting tumor Ki-67 response in this patient population. To our knowledge, this is the first study suggesting that serum TK1 activity may be a promising noninvasive pharmacodynamic marker of the antiproliferative effect of CDK4/6 inhibitors.
Uncontrolled cell proliferation is one of the key hallmarks of cancer [
26]. The value of TK1 as a cell proliferation marker was initially explored using IHC to study human breast cancers [
27]. Compared with the expression of proliferating cell nuclear antigen (PCNA), although expression of both TK1 and PCNA was significantly higher in malignant than in nonmalignant lesions, only TK1 was associated with tumor stage or histological grade [
28], suggesting that it is a better proliferative marker than PCNA. Serum TK1 was subsequently investigated as a tumor marker using monoclonal or polyclonal antibodies against TK1, demonstrating significantly higher levels in preoperative breast cancer patients than in healthy volunteers or patients with benign tumors or following curative surgery for breast cancer [
29]. A separate study demonstrated that higher serum TK1 levels 3 months after breast cancer surgery were associated with increased risk of both locoregional and distant recurrence [
30]. These earlier studies led to further interest in developing serum TK1 assays and investigation into clinical application. The innovative technology of the DiviTum assay enables the measurement of serum TK1 activity with high sensitivity and is compared favorably with other assay platforms [
6]. DiviTum serum TK1 activity has been explored as a prognostic marker in solid tumors, including breast cancer [
6‐
9,
21,
31]. Specifically, in a study of 368 women, including 149 healthy blood donors (control), 59 patients with benign breast disease (BBD), and 160 patients with primary breast cancer, serum TK1 activity was significantly higher in those with invasive breast cancer or with proliferative BBD than in those with nonproliferative BBD and healthy control subjects [
6]. Furthermore, serum TK1 activity was significantly associated with tumor size, lack of ER and PgR, tumor grade, and molecular subtype [
6]. Multivariate analyses adjusting for stage, grade, and HR status demonstrated that serum TK1 was an independent predictor of disease recurrence (
p = 0.013) [
6]. Additional studies demonstrated that TK1 activity was associated with progression-free survival and overall survival in patients with advanced and metastatic breast cancer [
21]. However, in few studies only have researchers investigated the clinical utilityof serum TK1 activity in monitoring therapeutic response to anti-neoplastic agents.
The particular interest in assessing serum TK1 activity in response to CDK4/6 inhibitors stems from the known cell cycle-inhibitory properties of these agents and their importance in the management of patients with advanced HR+ breast cancer [
18,
32]. Three CDK4/6 inhibitors, including palbociclib, ribociclib, and abemaciclib, are in clinical development for the treatment of breast cancer and other solid malignancies. All three CDK4/6 inhibitors have received U.S. Food and Drug Administration approval for the treatment of advanced HR+/HER2− breast cancer (Drugs@FDA;
https://www.accessdata.fda.gov/scripts/cder/daf/). However, despite this success, especially in endocrine-naive disease, resistance to CDK4/6 inhibitors eventually develops, and a significant proportion of patients with HR+/HER2− breast cancer fail to respond to CDK4/6 inhibitors in the second-line setting (>30%) or beyond (>60%) following progression on endocrine therapy [
11,
13,
33]. The clinical application of early response marker could lead to beneficial changes in determining the optimal therapeutic approach. Thus, the association between changes in serum TK1 activity and tumor Ki-67 response to palbociclib observed in the NeoPalAna trial provides the foundation for future exploration of the potential predictive nature of serum TK1 activity on response and progression-free survival in patients receiving CDK4/6 inhibitors for metastatic disease.
Our study is limited by its small sample size. However, the study is unique in its ability to obtain concurrent tumor biopsies and serum sample collections for Ki-67 IHC and TK1 activity at serial time points in patients with newly diagnosed, untreated HR+/HER2− breast cancer. The initial treatment with anastrozole monotherapy prior to the addition of palbociclib and the washout of palbociclib prior to surgery allows for the evaluation of dynamic changes in serum TK1 activity.
In this study, serum TK1 activity was not significantly changed following treatment with 28 days of anastrozole (Table
2). This is in spite of the statistically significant reduction in tumor TK1 mRNA at C1D1 (Table
2). In addition, a rise in serum TK1 activity at C1D1 was observed in 3 of the 11 patients in the anastrozole-sensitive group, although the changes were relatively small (from baseline of < 20, 38, and 44 Du/L to 34, 60, and 50 Du/L on C1D1, respectively) (Fig.
4). This was unexpected because anastrozole inhibits CDK4/6 indirectly through regulating cyclin D1 [
34,
35]. As demonstrated in our initial publication of the NeoPalAna trial, anastrozole regulated the mRNA expression levels of a wide range of genes, including those that were further suppressed by the addition of palbociclib [
19]. This is supported by the significant reduction in tumor TK1 mRNA following anastrozole at C1D1 and a reduction of serum TK1 activity in ~ 50% of patients. One possible explanation for the lack of significant serum TK1 response following anastrozole is the relatively significant contribution of serum TK1 activity from noncancer cells in the early-stage breast cancer setting. In a study in which investigators compared serum TK1 activity between healthy subjects and patients with breast cancer, the median and IQR value for TK1 activity activity was 16 Du/L (IQR 9–33 Du/L) in the healthy blood donors (
n = 149) and moderately increased in patients with primary breast cancer prior to surgical exicision (n=160), with median level of 37 Du/L (IQR 20–92 Du/L), respectively. [
6]. Although a statistically significant difference in serum TK1 activity was observed between the two groups, overlapping values exist. As shown in Table
2, the median baseline serum TK1 in patients enrolled in the NeoPalAna trial was 46 Du/L (IQR 25–73 Du/L), which is similar to that observed in the previous study [
6]. Therefore, the antitumor effect of anastrozole may be difficult to translate into serum TK1 changes, because the antiproliferative effect of anastrozole is restricted to the estrogen-dependent breast cancer cells, whereas the nonmalignant cells are otherwise not affected. This could particularly be the case in patients with lower baseline serum TK1 levels. The slight or moderate rise of serum TK1 activity at C1D1 in those three patients could also reflect fluctuations in individuals because TK1 has a short half-life [
4]. Therefore, studies including patients with advanced disease in which tumor cells contribute to the majority of the serum TK1 activity are warranted.
Similarly, the discordance between serum TK1 and tumor Ki-67 change in response to palbociclib, as well as, the low serum TK1 activity despite persistent tumor cell proliferation observed in patients in the palbociclib-resistant category, could be explained by the inhibitory effect of palbociclib on CDK4/6 in both cancer and non-cancer cells in this setting of early-stage disease. However, we could not rule out the possibility that serum TK1 activity is reduced only in CDK4/6-dependent cancer cells. Larger studies in patients with advanced disease will ultimately provide further insight and address these possibilities.
Acknowledgements
The authors thank the patients and families who participated in the NeoPalAna trial and the staff who cared for these patients. We thank the Siteman Cancer Center, the Tissue Procurement Core, the Biostatistics Core, the McDonnell Genome Institute, the Genomic and Pathology Service, and the Anatomical and Molecular Pathology Laboratory. The NeoPalAna trial was funded by Siteman Cancer Center grant P30 CA91842(to Timothy Eberlein, Section of Endocrine and Oncologic Surgery, Department of Surgery, Washington University School of Medicine, St. Louis, Missouri), the St. Louis Men’s Group Against Cancer (to CXM), a Susan G. Komen Promise Grant (to ME), a National Cancer Institute (NCI) Cancer Clinical Investigator Team Leadership Award (to CXM), and Pfizer (to CXM). CXM is a recipient of the NCI Clinical Investigator Team Leadership Award. ME is a McNair Medical Foundation Scholar and the recipient of a Cancer Prevention Research Institute of Texas senior investigator award. This work has been presented, in part, at the 2016 San Antonio Breast Cancer Symposium (abstract P5-04-02) and the 2017 American Association for Cancer Research Annual Meeting (abstract 2340).