PP1, PKA and DARPP‐32 in breast cancer: A retrospective assessment of protein and mRNA expression

Abstract Cyclic AMP–dependent protein kinase A (PKA) and protein phosphatase 1 (PP1) are proteins involved in numerous essential signalling pathways that modulate physiological and pathological functions. Both PP1 and PKA can be inhibited by dopamine‐ and cAMP‐regulated phosphoprotein 32 kD (DARPP‐32). Using immunohistochemistry, PKA and PP1 expression was determined in a large primary breast tumour cohort to evaluate associations between clinical outcome and clinicopathological criteria (n > 1100). In addition, mRNA expression of PKA and PP1 subunits was assessed in the METABRIC data set (n = 1980). Low protein expression of PKA was significantly associated with adverse survival of breast cancer patients; interestingly, this relationship was stronger in ER‐positive breast cancer patients. PP1 protein expression was not associated with patient survival. PKA and PP1 subunit mRNA was also assessed; PPP1CA, PRKACG and PRKAR1B were associated with breast cancer–specific survival. In patients with high expression of DARPP‐32, low expression of PP1 was associated with adverse survival when compared to high expression in the same group. PKA expression and PP1 expression are of significant interest in cancer as they are involved in a wide array of cellular processes, and these data indicate PKA and PP1 may play an important role in patient outcome.

define basal-like and HER2-positive breast cancers with a worse clinical outcome, 3 and PKA activity has also been implicated in HER2 resistance in vitro, in particular down-regulation of the PKA-RIIα subunit. 4 Like PKA, PP1 is a multifunctional protein involved in a large array of important cellular pathways. PP1 is an oligomeric enzyme that contains one catalytic subunit (encoded by PPP1CA, PPP1CB and PPP1CC) and at least one regulatory subunit (16 regulatory subunits). Importantly, PP1 plays a role in mitotic regulation, 5 glycogen metabolism and RNA splicing. In cancer, PP1 has been shown to play a role in the tumour microenvironment 6 and in steering cellular migration, 7 and it has been shown to interact with BRCA1. 8 Both PP1 and PKA can be inhibited by dopamine-and cAMPregulated phosphoprotein 32 kD (DARPP-32), also known as protein phosphatase one regulatory subunit 1B and encoded by PPP1R1B. DARPP-32 was originally described as a signalling protein highly concentrated in regions of the brain enriched in dopaminergic nerve terminals. 9 DARPP-32 is able to inhibit both PKA and PP1 dependent upon its phosphorylation status. For instance, DARPP-32 is phosphorylated on threonine 34 by PKA to allow it to act as a potent inhibitor of PP1. 10 Cyclin-dependent kinase (Cdk)-5 can phosphorylate DARPP-32 on threonine 75 to allow it to act as a PKA inhibitor. 11 In a variety of cancers, a truncated DARPP-32 splice variant (t-DARPP) is expressed; importantly, this isoform lacks the ability to inhibit PP1, but retains PKA inhibitory activity. 12 Phosphorylation of t-DARPP has been shown to mediate PKA activation, which appears to be via association between t-DARPP and the regulatory R1 subunit. 13 There is accumulating evidence that differential expression of DARPP-32 and t-DARPP is functionally significant in numerous tumour types (reviewed in 14 ). Low levels of DARPP-32 protein expression are associated with shorter cancer-specific survival of breast and ovarian cancer patients. 15,16 In trastuzumab-resistant breast cancer cells, t-DARPP activated IGF-1R signalling, which stimulated glycolysis and conferred trastuzumab resistance. 17 In gastric cancer, DARPP-32 is frequently overexpressed in the early stages of gastric cancer. 18 In lung cancer, DARPP-32 and t-DARPP promoted tumour cell growth in vivo and in vitro. 19,20 In pancreatic cancer, loss of HIF1α increased PPP1R1B expression, and DARPP-32 promoted degradation of p53 through phosphorylation of MDM2. 21 DARPP-32 is phosphorylated by PKA or Cdk5 to alter its inhibitory activity; Cdk5 is involved in neuronal maturation but is also implicated in cancer and neurodegenerative disorders. In breast cancer, low Cdk5 expression has also been shown to be associated with adverse survival of breast cancer patients. 16 In breast cancer, we hypothesize that a reduction in DARPP-32 and an increase in t-DARPP in breast cancer result in a concomitant alteration of PKA and PP1 signalling. We sought to determine the expression of PKA and PP-1 mRNA and protein in a large cohort of early-stage invasive breast cancer patients to understand their relationship between survival and clinicopathological criteria and to determine their relationship between DARPP-32 and Cdk5 expression.

| Immunohistochemistry
Immunohistochemistry was performed on tissue microarrays comprised of single 0.6 mm cores taken from a representative tumour area as assessed on haematoxylin and eosin-stained sections by a specialist breast cancer histopathologist. Immunohistochemical staining was achieved using a Novolink Polymer Detection Kit (Leica) according to the manufacturers' instructions, the use of which has been described previously. Briefly, tissue was deparaffinized in xylene and rehydrated in ethanol followed by water.

Statistical analysis was performed using IBM SPSS Statistics
(version 24). X-tile was used to determine cut-points for assessment using breast cancer-specific survival. 23 Single measure intraclass correlation coefficients were used to determine concordance between scorers with levels above 0.7 that indicated good concordance. The Pearson chi-squared test of association was used to determine the relationship between categorized protein expression and clinicopathological variables. Survival curves were plotted according to the Kaplan-Meier method with significance determined using the log-rank test. Multivariate survival analysis was computed using the Cox regression analysis. All differences were deemed statistically significant at the level of P ≤ .05. This study is reported according to REMARK criteria. 24 Broad Institute Morpheus software was used to visualize data (https://softw are. broad insti tute.org/morpheus). Expression of PKA and PP1 was F I G U R E 1 Heat map of DARPP-32, PP1, PKA and Cdk5 expression (A). Representative photomicrographs of high (B) and low (C) PKA expression, and high (D) and low (E) PP1 expression are shown at 10× magnification with a 20× magnification inset box. Scale bar represents 100 µM. Similarity matrix of Pearson's correlation coefficients between protein expression (F) TA B L E 1 Associations between the cytoplasmic and nuclear expression of DARPP-32, determined in the discovery cohort and validation cohort using immunohistochemistry, with clinicopathological variables

| PKA and PP1 protein staining and frequency and correlation of expression
The majority of PKA expression was cytoplasmic with weak to strong staining intensity observed in tumour cells; representative staining is shown in Figure 1B Representative PKA and PP1 protein expression is demonstrated in Figure 1A.
Correlations between the expression of PKA and PP1 together with DARPP-32 and Cdk5 were assessed. Many of the combinations were statistically significant; however, they were of weak biological significance ( Figure 1E). PKA expression was signifi-
PKA expression was not associated with any other clinicopathological variables (Table 1).

| Association between PKA and PP1 protein expression and survival
Low PKA expression was significantly associated with adverse survival of breast cancer patients (P = .040) (Figure 2A)  Figure 3B), low PP1 and low DARPP-32 (P = .001; Figure 3C) and combination of low PP1 and low Cdk5 (P = .010; Figure 3D) showed significant association with shorter survival.
However, the combined expression of PKA and DARPP-32 was not associated with breast cancer-specific survival (P = .139) ( Figure 3A).

| PKA and PP1 mRNA expression in the METABRIC cohort
PKA and PP1 mRNA expression was assessed for association with patient survival in the METABRIC cohort using probe-level information. Probe-level information was used so that correlations with PPP1R1B variants could be assessed.  Figure 4A,B.   and PRKAR2B has been observed, in combination with an observed increase with PPP1R1B. 4 Our findings suggest that PRKAR1B and PRKACG were associated with disease-specific survival of the total patient cohort.

| D ISCUSS I ON
Our findings suggest that PPP1CA, PRKACG and PRKAR1B are associated with breast cancer-specific survival. Previously published studies have indicated a role for PPP1CA in cancer 26,27 ; in breast cancer, PPP1CA may play a role in cancer stem cell populations. 28 Although a role for PRKACG and PRKAR1B in breast cancer has not been described, PRKACG has been shown to act as a RUNX1mutation associated hub gene in acute myeloid leukaemia 29 and single nucleotide polymorphisms of PRKAR1B are associated with inferior survival of advanced renal cell cancer patients. 30 In previous studies, we demonstrated that low DARPP-32 and Cdk5 expression is associated with adverse survival of breast cancer patients. 15,16 We hypothesized that a reduction in DARPP-32

We thank the Nottingham Health Science Biobank and Breast
Cancer Now for the provision of tissue samples.

DATA AVA I L A B I L I T Y S TAT E M E N T
The METABRIC data set is publically available at https://www.ebi.
ac.uk/ega/studi es/EGAS0 00000 00098. Immunohistochemistry data sets analysed during the current study are available from the corresponding author on reasonable request.