Dopamine and cAMP‐regulated phosphoprotein 32kDa (DARPP‐32), protein phosphatase‐1 and cyclin‐dependent kinase 5 expression in ovarian cancer

Abstract Dopamine and cyclic‐AMP activated phosphoprotein Mr32kDa (DARPP‐32) is a central signalling protein in neurotransmission. Following DARPP‐32 phosphorylation by protein kinase A (PKA), DARPP‐32 becomes a potent protein phosphatase 1 (PP1) inhibitor. DARPP‐32 can itself inhibit PKA following DARPP‐32 phosphorylation by cyclin‐dependent kinase 5 (Cdk5). Increasing evidence indicates a role for DARPP‐32 and its associated signalling pathways in cancer; however, its role in ovarian cancer remains unclear. Using immunohistochemistry, expression of DARPP‐32, PP1 and Cdk5 was determined in a large cohort of primary tumours from ovarian cancer patients (n = 428, 445 and 434 respectively) to evaluate associations between clinical outcome and clinicopathological criteria. Low cytoplasmic and nuclear DARPP‐32 expression was associated with shorter patient overall survival and progression‐free survival (P = .001, .001, .004 and .037 respectively). Low nuclear and cytoplasmic DARPP‐32 expression remained significantly associated with overall survival in multivariate Cox regression (P = .045, hazard ratio (HR) = 0.734, 95% confidence interval (CI) = 0.542‐0.993 and P = .001, HR = 0.494, 95% CI = 0.325‐0.749, respectively). High cytoplasmic and nuclear PP1 expression was associated with shorter patient overall survival and high cytoplasmic PP1 expression with shorter progression‐free survival (P = .005, .033, and .037, respectively). High Cdk5 expression was associated with shorter progression‐free survival (P = .006). These data suggest a role for DARPP‐32 and associated signalling kinases as prognostic markers with clinical utility in ovarian cancer.


| INTRODUC TI ON
Dopamine and cyclic-AMP (cAMP)-regulated phosphoprotein Mr32kDa (DARPP-32) is a target of dopamine and cAMP and is found enriched in dopaminoceptive nerve terminals where it acts as a central signalling molecule. Dependent upon its phosphorylation state, DARPP-32 has been shown to act as a phosphatase inhibitor or a kinase. DARPP-32 acts as an inhibitor of protein phosphatase 1 (PP1) and protein kinase A (PKA). 1 Phosphorylation at Threonine (Thr)-34 by PKA converts DARPP-32 to a potent inhibitor of PP1, whereas phosphorylation at Thr-75 by cyclin-dependent kinase 5 (Cdk5) or cell division cycle (cdc)2 converts DARPP-32 into a PKA inhibitor. 2 There is further complexity in this signalling, with phosphorylation of DARPP-32 Serine (Ser)-137 by casein kinase 1 (CK1) able to prevent dephosphorylation of Thr-34. 3 A truncated DARPP-32 splice variant, t-DARPP, exists that TA B L E 1 Associations between the cytoplasmic and nuclear expression of DARPP-32, PP1 and Cdk5 determined using immunohistochemistry with clinicopathological variables. The P values are resultant from Pearson's χ 2 test of association, and significant values (P ≤ .05) are highlighted in bold to function as a PKA inhibitor. Cdk5 is considered a neuronal serine/ threonine kinase that is predominantly activated by p35 or p39; its function has been implicated in a number of tumourigenic pathways, with a number of important substrates in addition to DARPP-32, including p53 and AKT (reviewed in Ref. [14]). Studies in a number of tumour types have demonstrated that high Cdk5 expression is associated with clinicopathological criteria associated with poor prognosis, and in some cases, shortened disease-specific survival itself [15][16][17] ; however, the reverse has been observed in gastric cancer. 18,19 There is increasing evidence that DARPP-32, PP1 and Cdk5 have a role in various tumour types; however, expression of DARPP-32 and PP1 has not previously been described in ovarian cancer, although DARPP-32 has been implicated in follicular development. 20 In ovarian cancer, in vitro studies have indicated a role for Cdk5 in paclitaxel sensitivity, 21

| Patient cohorts
Patients received treatment at Nottingham University Hospitals be-

| Immunohistochemistry
Immunohistochemical staining was conducted using a Novolink Polymer Detection kit (Leica) on a tissue microarray comprised of single 0.6mm cores from 575 ovarian tumours taken from a representative area as assessed by a specialist ovarian cancer histopathologist; the use of which has been described previously. [27][28][29] Staining was conducted according to manufacturers' instructions and has been described previously. 30

| Relationship between DARPP-32, PP1 and Cdk5
The relationship between DARPP-32, PP1 and Cdk5 cytoplasmic and nuclear expression was explored using the Spearman rank correlation coefficient. DARPP-32 cytoplasmic expression was significantly correlated with DARPP-32 nuclear expression (r 2 = .858, P < .001). PP1 cytoplasmic expression was significantly correlated with PP1 nuclear expression (r 2 = .487, P < .001). Cdk5 cytoplasmic expression was significantly correlated with Cdk5 nuclear expression (r 2 = .615, P < .001). All of these correlations indicate a strong biological relationship between nuclear and cytoplasmic expression of each protein that has not been assessed further within this study.

| Association between DARPP-32, PP1 and Cdk5 protein expression and overall survival
Low DARPP-32 expression in the cytoplasm and in the nucleus was associated with adverse survival (both P = .001; Figure 2A Cdk5 cytoplasmic and nuclear expression was not associated with patient survival (Figure 2E,F).

| Association between DARPP-32, PP1 and Cdk5 protein expression and progression-free survival
Low DARPP-32 expression in the cytoplasm and in the nucleus was also associated with shorter progression-free survival (P = .004 and .037, respectively) ( Figure 3A,B). Low cytoplasmic DARPP-32

| Association between DARPP-32, PP1 and Cdk5 protein expression combinations and overall survival
High and low cytoplasmic DARPP-32 expression in high and low expression groups of cytoplasmic PP1 and Cdk5 were assessed to understand the impact of these proteins on patient survival. The combination of low DARPP-32 expression and high PP1 expression was associated with shorter overall survival (P < .001; Figure 4A).
Low DARPP-32 and high Cdk5 expression was associated with shorter overall survival (P < .001; Figure 4B).  In addition to overall survival, we were also able to test associations with progression-free survival of ovarian cancer patients. Low cytoplasmic and nuclear DARPP-32 expression, high cytoplasmic PP1 expression and high cytoplasmic Cdk5 expression were associated with shorter progression-free survival. As none of the proteins were associated with response to chemotherapy this suggests that the association with progression-free survival is not linked with altered sensitivity to chemotherapy.

| D ISCUSS I ON
Finally, low expression of DARPP-32 in tumours with high expression of Cdk5 or PP1 was more strongly associated with shorter survival that the alternative combinations. These combined results suggest a loss of DARPP-32 and/or protein function may be important in ovarian cancer, in particular those with high expression levels of PP1 or Cdk5.

| CON CLUS ION
Low cytoplasmic and nuclear DARPP-32 expression and high cytoplasmic and nuclear PP1 expression are associated with shorter survival in ovarian cancer patients. Importantly, both cytoplasmic and nuclear expression of DARPP-32 remain associated with overall survival when other confounding factors are included in multivariate analysis. In addition, low cytoplasmic and nuclear DARPP-32, high cytoplasmic PP1 and high cytoplasmic Cdk5 expression is associated with adverse progression-free survival.
These findings warrant further investigation in larger patient cohorts but indicate that DARPP-32 expression may be of clinical relevance in ovarian cancer.

ACK N OWLED G EM ENTS
SJS is funded through the University of Nottingham's Research Vision as a Nottingham Research Fellow and previously through the University of Nottingham Life Cycle 6 Campaign.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

AUTH O R CO NTR I B UTI O N
SS, SZ, SY and BS conducted the studies and collected the data; SD provided tissue; SS conducted statistical analysis; SGM and SS conceived the study; SS wrote the manuscript; and all authors approved the manuscript for submission.

DATA AVA I L A B I L I T Y S TAT E M E N T
Immunohistochemistry data sets analysed during the current study are available from the corresponding author by request.