Oncogenic roles of PRL-3 in FLT3-ITD induced acute myeloid leukaemia

FLT3-ITD mutations are prevalent mutations in acute myeloid leukaemia (AML). PRL-3, a metastasis-associated phosphatase, is a downstream target of FLT3-ITD. This study investigates the regulation and function of PRL-3 in leukaemia cell lines and AML patients associated with FLT3-ITD mutations. PRL-3 expression is upregulated by the FLT3-STAT5 signalling pathway in leukaemia cells, leading an activation of AP-1 transcription factors via ERK and JNK pathways. PRL-3-depleted AML cells showed a significant decrease in cell growth. Clinically, high PRL-3 mRNA expression was associated with FLT3-ITD mutations in four independent AML datasets with 1158 patients. Multivariable Cox-regression analysis on our Cohort 1 with 221 patients identified PRL-3 as a novel prognostic marker independent of other clinical parameters. Kaplan-Meier analysis showed high PRL-3 mRNA expression was significantly associated with poorer survival among 491 patients with normal karyotype. Targeting PRL-3 reversed the oncogenic effects in FLT3-ITD AML models in vitro and in vivo. Herein, we suggest that PRL-3 could serve as a prognostic marker to predict poorer survival and as a promising novel therapeutic target for AML patients.

Metastasis-associated PRL-3 induces EGFR activation and addiction in cancer cells

Metastasis-associated phosphatase of regenerating liver-3 (PRL-3) has pleiotropic effects in driving cancer progression, yet the signaling mechanisms of PRL-3 are still not fully understood. Here, we provide evidence for PRL-3-induced hyperactivation of EGFR and its downstream signaling cascades in multiple human cancer cell lines. Mechanistically, PRL-3-induced activation of EGFR was attributed primarily to transcriptional downregulation of protein tyrosine phosphatase 1B (PTP1B), an inhibitory phosphatase for EGFR. Functionally, PRL-3-induced hyperactivation of EGFR correlated with increased cell growth, promigratory characteristics, and tumorigenicity. Moreover, PRL-3 induced cellular addiction to EGFR signaling, as evidenced by the pronounced reversion of these oncogenic attributes upon EGFR-specific inhibition. Of clinical significance, we verified elevated PRL-3 expression as a predictive marker for favorable therapeutic response in a heterogeneous colorectal cancer (CRC) patient cohort treated with the clinically approved anti-EGFR antibody cetuximab. The identification of PRL-3-driven EGFR hyperactivation and consequential addiction to EGFR signaling opens new avenues for inhibiting PRL-3-driven cancer progression. We propose that elevated PRL-3 expression is an important clinical predictive biomarker for favorable anti-EGFR cancer therapy.

PRL-3 activates mTORC1 in Cancer Progression

PRL-3, a metastasis-associated phosphatase, is known to exert its oncogenic functions through activation of PI3K/Akt, which is a key regulator of the rapamycin-sensitive mTOR complex 1 (mTORC1), but a coherent link between PRL-3 and activation of mTOR has not yet been formally demonstrated. We report a positive correlation between PRL-3 expression and mTOR phospho-activation in clinical tumour samples and mouse models of cancer and demonstrate that PRL-3 increased downstream signalling to the mTOR substrates, p70S6K and 4E-BP1, by increasing PI3K/Akt-mediated activation of Rheb-GTP via TSC2 suppression. We also show that PRL-3 increases mTOR translocation to lysosomes via increased mTOR binding affinity to Rag GTPases in an Akt-independent manner, demonstrating a previously undescribed mechanism of action for PRL-3. PRL-3 also enhanced matrix metalloproteinase-2 secretion and cellular invasiveness via activation of mTOR, attributes which were sensitive to rapamycin treatment. The downstream effects of PRL-3 were maintained even under conditions of environmental stress, suggesting that PRL-3 provides a strategic survival advantage to tumour cells via its effects on mTOR.

A role of autophagy in PTP4A3-driven cancer progression

Autophagy, a "self-eating" cellular process, has dual roles in promoting and suppressing tumor growth, depending on cellular context. PTP4A3/PRL-3, a plasma membrane and endosomal phosphatase, promotes multiple oncogenic processes including cell proliferation, invasion, and cancer metastasis. In this study, we demonstrate that PTP4A3 accumulates in autophagosomes upon inhibition of autophagic degradation. Expression of PTP4A3 enhances PIK3C3-BECN1-dependent autophagosome formation and accelerates LC3-I to LC3-II conversion in an ATG5-dependent manner. PTP4A3 overexpression also enhances the degradation of SQSTM1, a key autophagy substrate. These functions of PTP4A3 are dependent on its catalytic activity and prenylation-dependent membrane association. These results suggest that PTP4A3 functions to promote canonical autophagy flux. Unexpectedly, following autophagy activation, PTP4A3 serves as a novel autophagic substrate, thereby establishing a negative feedback-loop that may be required to fine-tune autophagy activity. Functionally, PTP4A3 utilizes the autophagy pathway to promote cell growth, concomitant with the activation of AKT. Clinically, from the largest ovarian cancer data set (GSE 9899, n = 285) available in GEO, high levels of expression of both PTP4A3 and autophagy genes significantly predict poor prognosis of ovarian cancer patients. These studies reveal a critical role of autophagy in PTP4A3-driven cancer progression, suggesting that autophagy could be a potential Achilles heel to block PTP4A3-mediated tumor progression in stratified patients with high expression of both PTP4A3 and autophagy genes.

Biochemical characterization of Aprataxin, the protein deficient in Ataxia with Oculomotor Apraxia type 1

Neurodegenerative disorders are heterogenous in nature and include a range of ataxias with oculomotor apraxia, which are characterised by a wide variety of neurological and ophthalmological features. This family includes recessive and dominant disorders. A subfamily of autosomal recessive cerebellar ataxias are characterised by defects in the cellular response to DNA damage. These include the well characterised disorders Ataxia-Telangiectasia (A-T) and Ataxia-Telangiectasia Like Disorder (A-TLD) as well as the recently identified diseases Spinocerebellar ataxia with axonal neuropathy Type 1 (SCAN1), Ataxia with Oculomotor Apraxia Type 2 (AOA2), as well as the subject of this thesis, Ataxia with Oculomotor Apraxia Type 1 (AOA1). AOA1 is caused by mutations in the APTX gene, which is located at chromosomal locus 9p13. This gene codes for the 342 amino acid protein Aprataxin. Mutations in APTX cause destabilization of Aprataxin, thus AOA1 is a result of Aprataxin deficiency. Aprataxin has three functional domains, an N-terminal Forkhead Associated (FHA) phosphoprotein interaction domain, a central Histidine Triad (HIT) nucleotide hydrolase domain and a C-terminal C2H2 zinc finger. Aprataxins FHA domain has homology to FHA domain of the DNA repair protein 5’ polynucleotide kinase 3’ phosphatase (PNKP). PNKP interacts with a range of DNA repair proteins via its FHA domain and plays a critical role in processing damaged DNA termini. The presence of this domain with a nucleotide hydrolase domain and a DNA binding motif implicated that Aprataxin may be involved in DNA repair and that AOA1 may be caused by a DNA repair deficit. This was substantiated by the interaction of Aprataxin with proteins involved in the repair of both single and double strand DNA breaks (XRay Cross-Complementing 1, XRCC4 and Poly-ADP Ribose Polymerase-1) and the hypersensitivity of AOA1 patient cell lines to single and double strand break inducing agents. At the commencement of this study little was known about the in vitro and in vivo properties of Aprataxin. Initially this study focused on generation of recombinant Aprataxin proteins to facilitate examination of the in vitro properties of Aprataxin. Using recombinant Aprataxin proteins I found that Aprataxin binds to double stranded DNA. Consistent with a role for Aprataxin as a DNA repair enzyme, this binding is not sequence specific. I also report that the HIT domain of Aprataxin hydrolyses adenosine derivatives and interestingly found that this activity is competitively inhibited by DNA. This provided initial evidence that DNA binds to the HIT domain of Aprataxin. The interaction of DNA with the nucleotide hydrolase domain of Aprataxin provided initial evidence that Aprataxin may be a DNA-processing factor. Following these studies, Aprataxin was found to hydrolyse 5’adenylated DNA, which can be generated by unscheduled ligation at DNA breaks with non-standard termini. I found that cell extracts from AOA1 patients do not have DNA-adenylate hydrolase activity indicating that Aprataxin is the only DNA-adenylate hydrolase in mammalian cells. I further characterised this activity by examining the contribution of the zinc finger and FHA domains to DNA-adenylate hydrolysis by the HIT domain. I found that deletion of the zinc finger ablated the activity of the HIT domain against adenylated DNA, indicating that the zinc finger may be required for the formation of a stable enzyme-substrate complex. Deletion of the FHA domain stimulated DNA-adenylate hydrolysis, which indicated that the activity of the HIT domain may be regulated by the FHA domain. Given that the FHA domain is involved in protein-protein interactions I propose that the activity of Aprataxins HIT domain may be regulated by proteins which interact with its FHA domain. We examined this possibility by measuring the DNA-adenylate hydrolase activity of extracts from cells deficient for the Aprataxin-interacting DNA repair proteins XRCC1 and PARP-1. XRCC1 deficiency did not affect Aprataxin activity but I found that Aprataxin is destabilized in the absence of PARP-1, resulting in a deficiency of DNA-adenylate hydrolase activity in PARP-1 knockout cells. This implies a critical role for PARP-1 in the stabilization of Aprataxin. Conversely I found that PARP-1 is destabilized in the absence of Aprataxin. PARP-1 is a central player in a number of DNA repair mechanisms and this implies that not only do AOA1 cells lack Aprataxin, they may also have defects in PARP-1 dependant cellular functions. Based on this I identified a defect in a PARP-1 dependant DNA repair mechanism in AOA1 cells. Additionally, I identified elevated levels of oxidized DNA in AOA1 cells, which is indicative of a defect in Base Excision Repair (BER). I attribute this to the reduced level of the BER protein Apurinic Endonuclease 1 (APE1) I identified in Aprataxin deficient cells. This study has identified and characterised multiple DNA repair defects in AOA1 cells, indicating that Aprataxin deficiency has far-reaching cellular consequences. Consistent with the literature, I show that Aprataxin is a nuclear protein with nucleoplasmic and nucleolar distribution. Previous studies have shown that Aprataxin interacts with the nucleolar rRNA processing factor nucleolin and that AOA1 cells appear to have a mild defect in rRNA synthesis. Given the nucleolar localization of Aprataxin I examined the protein-protein interactions of Aprataxin and found that Aprataxin interacts with a number of rRNA transcription and processing factors. Based on this and the nucleolar localization of Aprataxin I proposed that Aprataxin may have an alternative role in the nucleolus. I therefore examined the transcriptional activity of Aprataxin deficient cells using nucleotide analogue incorporation. I found that AOA1 cells do not display a defect in basal levels of RNA synthesis, however they display defective transcriptional responses to DNA damage. In summary, this thesis demonstrates that Aprataxin is a DNA repair enzyme responsible for the repair of adenylated DNA termini and that it is required for stabilization of at least two other DNA repair proteins. Thus not only do AOA1 cells have no Aprataxin protein or activity, they have additional deficiencies in PolyADP Ribose Polymerase-1 and Apurinic Endonuclease 1 dependant DNA repair mechanisms. I additionally demonstrate DNA-damage inducible transcriptional defects in AOA1 cells, indicating that Aprataxin deficiency confers a broad range of cellular defects and highlighting the complexity of the cellular response to DNA damage and the multiple defects which result from Aprataxin deficiency. My detailed characterization of the cellular consequences of Aprataxin deficiency provides an important contribution to our understanding of interlinking DNA repair processes.

Pten and Ndufb8 aberrations in cervical cancer tissue

Cervical cancer is one of the world's major health issues. Despite many studies in this field, the carcinogenetic events of malignant conversion in cervical tumours have not been significantly characterised. The first aim of this project was to investigate the mutation status of the tumour suppressor gene- Phosphatase and Tension Homolog (PTEN)- in cervical cancer tissue. The second aim of this study was the analysis in the same cervical cancer tissue for aberrations in the mitochondrial electron transport chain subunit gene NDUFB8, which is localised to the same chromosomal contig as PTEN. The third aim was the evaluation of the potential therapeutic anti-cancer drug 2,4-Thiazolidinediones (TZDs) and its affect in regulating the PTEN protein in a cervical cancer cell line (HeLa). To approach the aims, paraffin-embedded cancerous cervical tissue and non-cancerous cervical tissue were obtained. DNA recovered from those tissues was then used to investigate the putative genomic changes regarding the NDUFB8 gene utilising SYBR Green I Real-Time PCR. The PTEN gene was studied via Dual-Labelled probe Real-Time PCR. To investigate the protein expression change of the PTEN protein, HeLa cells were firstly treated with different concentrations of 2,4-Thiazolidinediones and the level of PTEN protein expression was then observed utilising standard protein assays. Results indicated that there were putative copy-number changes between the cancerous cervical tissue and non-cancerous cervical tissue, with regard to the PTEN locus. This implies a potential gain of the PTEN gene in cancerous cervical tissue. With regards to normal cervical tissue versus cancerous cervical tissue no significant melting temperature differences were observed with the SYBR Green I Real-Time PCR in respect to the NDUFB8 gene. A putative up-regulation of PTEN protein was observed in TZD treated HeLa cells. © 2008 Springer Science+Business Media, LLC.

Accumulation of the SET protein in HEK293T cells and mild oxidative stress: cell survival or death signaling

SET protein (I2PP2A) is an inhibitor of PP2A, which regulates the phosphorylated Akt (protein kinase B) levels. We assessed the effects of SET overexpression in HEK293T cells, both in the presence and the absence of mild oxidative stress induced by 50 mu M tert-butyl hydroperoxide. Immunoblotting assays demonstrated that SET accumulated in HEK293T cells and increased the levels of phosphorylated Akt and PTEN; in addition, SET decreased glutathione antioxidant defense of cell and increased expression of genes encoding antioxidant defense proteins. Immunofluorescence analysis demonstrated that accumulated SET was equally distributed in cytoplasm and nucleus; however, in cells that had been exposed to oxidative stress, SET was found in large aggregates in the cytoplasm. SET accumulation in HEK293T cells correlated with inhibition of basal apoptosis as evidenced by a decrease in annexin V staining and activity of caspases; under mild oxidative stress, SET accumulation correlated with caspase-independent cell death, as evidenced by increased PI and annexin V/PI double staining. The results suggest that accumulated SET could act via Akt/PTEN either as cell survival signal or as oxidative stress sensor for cell death.

Overexpression of a Rrp1 transgene reduces the somatic mutation and recombination frequency induced by oxidative DNA damage in Drosophila melanogaster.

Recombination repair protein 1 (Rrp1) includes a C-terminal region homologous to several DNA repair proteins, including Escherichia coli exonuclease III and human APE, that repair oxidative and alkylation damage to DNA. The nuclease activities of Rrp1 include apurinic/apyrimidinic endonuclease, 3'-phosphodiesterase, 3'-phosphatase, and 3'-exonuclease. As shown previously, the C-terminal nuclease region of Rrp1 is sufficient to repair oxidative- and alkylation-induced DNA damage in repair-deficient E. coli mutants. DNA strand-transfer and single-stranded DNA renaturation activities are associated with the unique N-terminal region of Rrp1, which suggests possible additional functions that include recombinational repair or homologous recombination. By using the Drosophila w/w+ mosaic eye system, which detects loss of heterozygosity as changes in eye pigmentation, somatic mutation and recombination frequencies were determined in transgenic flies overexpressing wild-type Rrp1 protein from a heat-shock-inducible transgene. A large decrease in mosaic clone frequency is observed when Rrp1 overexpression precedes treatment with gamma-rays, bleomycin, or paraquat. In contrast, Rrp1 overexpression does not alter the spot frequency after treatment with the alkylating agents methyl methanesulfonate or methyl nitrosourea. A reduction in mosaic clone frequency depends on the expression of the Rrp1 transgene and on the nature of the induced DNA damage. These data suggest a lesion-specific involvement of Rrp1 in the repair of oxidative DNA damage.

Functional genomic and molecular analyses of transcripts related to glycerol synthesis and trafficking in rainbow smelt (Osmerus mordax) using cold temperature-induced models of glycerol production

The rainbow smelt (Osmerus mordax) is an anadromous teleost that produces type II antifreeze protein (AFP) and accumulates modest urea and high glycerol levels in plasma and tissues as adaptive cryoprotectant mechanisms in sub-zero temperatures. It is known that glyceroneogenesis occurs in liver via a branch in glycolysis and gluconeogenesis and is activated by low temperature; however, the precise mechanisms of glycerol synthesis and trafficking in smelt remain to be elucidated. The objective of this thesis was to provide further insight using functional genomic techniques [e.g. suppression subtractive hybridization (SSH) cDNA library construction, microarray analyses] and molecular analyses [e.g. cloning, quantitative reverse transcription - polymerase chain reaction (QPCR)]. Novel molecular mechanisms related to glyceroneogenesis were deciphered by comparing the transcript expression profiles of glycerol (cold temperature) and non-glycerol (warm temperature) accumulating hepatocytes (Chapter 2) and livers from intact smelt (Chapter 3). Briefly, glycerol synthesis can be initiated from both amino acids and carbohydrate; however carbohydrate appears to be the preferred source when it is readily available. In glycerol accumulating hepatocytes, levels of the hepatic glucose transporter (GLUT2) plummeted and transcript levels of a suite of genes (PEPCK, MDH2, AAT2, GDH and AQP9) associated with the mobilization of amino acids to fuel glycerol synthesis were all transiently higher. In contrast, in glycerol accumulating livers from intact smelt, glycerol synthesis was primarily fuelled by glycogen degradation with higher PGM and PFK (glycolysis) transcript levels. Whether initiated from amino acids or carbohydrate, there were common metabolic underpinnings. Increased PDK2 (an inhibitor of PDH) transcript levels would direct pyruvate derived from amino acids and / or DHAP derived from G6P to glycerol as opposed to oxidation via the citric acid cycle. Robust LIPL (triglyceride catabolism) transcript levels would provide free fatty acids that could be oxidized to fuel ATP synthesis. Increased cGPDH (glyceroneogenesis) transcript levels were not required for increased glycerol production, suggesting that regulation is more likely by post-translational modification. Finally, levels of a transcript potentially encoding glycerol-3-phosphatase, an enzyme not yet characterized in any vertebrate species, were transiently higher. These comparisons also led to the novel discoveries that increased G6Pase (glucose synthesis) and increased GS (glutamine synthesis) transcript levels were part of the low temperature response in smelt. Glucose may provide increased colligative protection against freezing; whereas glutamine could serve to store nitrogen released from amino acid catabolism in a non-toxic form and / or be used to synthesize urea via purine synthesis-uricolysis. Novel key aspects of cryoprotectant osmolyte (glycerol and urea) trafficking were elucidated by cloning and characterizing three aquaglyceroporin (GLP)-encoding genes from smelt at the gene and cDNA levels in Chapter 4. GLPs are integral membrane proteins that facilitate passive movement of water, glycerol and urea across cellular membranes. The highlight was the discovery that AQP10ba transcript levels always increase in posterior kidney only at low temperature. This AQP10b gene paralogue may have evolved to aid in the reabsorption of urea from the proximal tubule. This research has contributed significantly to a general understanding of the cold adaptation response in smelt, and more specifically to the development of a working scenario for the mechanisms involved in glycerol synthesis and trafficking in this species.

Synthesis, Structure, and Phosphatase-Like Activity of a New Trinuclear Gd Complex with the Unsymmetrical Ligand H(3)L As a Model for Nucleases

The new trinuclear gadolinium complex [Gd(3)L(2)(NO(3))(2)(H(2)O)(4)]NO(3)center dot 8H(2)O (1) with the unsymmetrical ligand 2-[N-bis-(2-pyridylmethyl)aminomethyl]-4-methyl-6-[N-bis(2-hydroxy-2-oxoethyl)aminomethyl] phenol (H(3)L) was synthesized and characterized. The new ligand H(3)L was obtained in good yield. Complex I crystallizes in an orthorhombic cell, space group Pcab. Kinetic studies show that complex 1 is highly active in the hydrolysis of the substrate 2,4-bis(dinitrophenyl)phosphate (K(m) = 4.09 mM, V(max) = 2.68 x 10(-2) mM s(-1), and k(cat) = V(max)/[1] = 0.67 s(-1)). Through a potentiometric study and determination of the kinetic behavior of 1 in acetonitrile/water solution, the species present in solution could be identified, and a trinuclear monohydroxo species appears to be the most prominent catalyst under mild conditions. Complex 1 displays high efficiency in DNA hydrolytic cleavage, and complete kinetic studies were carried out (K(m) = 4.57 x 10(-4) M, K(cat)` = 3.42 h(-1), and k(cat)`/K(m) = 7.48 x 10(3) M(-1) h(-1)). Studies with a radical scavenger (dimethyl sulfoxide, DMSO) showed that it did not inhibit the activity, indicating the hydrolytic action of 1 in the cleavage of DNA, and studies on the incubation of distamycin with plasmid DNA suggest that 1 is regio-specific, interacting with the minor groove of DNA.

Purification and properties of acid phosphatase (EC 3.1.3.2) secreted by strain 74A of the mould Neurospora crassa

Both P-i-repressible acid phosphatases, IIb (mycelial) and IIc (extracellular), synthesized by Neurospora crassa and purified to apparent homogeneity by 7.5% PAGE, are monomers, are inhibited by 2 mM ZnCl2 and are nonspecifically stimulated by salts. However, the IIc form is activated by p-nitrophenylphosphate (in a negative cooperativity effect with a K-0.5 of 2.5 mM) whereas form IIb shows Michaelis kinetics, with a K-m of 0.5 mM. Thus, since both enzymatic forms may be expressed by the same gene (pho-3), it is possible that post-translational modifications lead to the excretion of an enzymatic form with altered Michaelis kinetics compared with the enzymatic form retained by the mycelium.

THE PHO-2A MUTANT OF NEUROSPORA-CRASSA WHICH IS DEFICIENT IN PI-REPRESSIBLE ALKALINE-PHOSPHATASE (EC 3.1.3.1) IS ALSO DEFECTIVE IN PI-REPRESSIBLE ACID-PHOSPHATASE (EC 3.1.3.2)

1. The mycelial Pi-repressible acid phosphatase presented p-nitrophenylphosphatase activity with negative cooperativity and Michaelian behavior when synthesized by the wild-type and pho-2A mutant strains of Neurospora crassa, respectively.2. The major acid phosphatase present in cell extracts of the pho-2A mutant of N. crassa grown in low Pi medium is more thermolabile (t1/2 = 4 min at 54-degrees-C, pH 5.4) than that of the wild strain (stable for at least 80 min at 54-degrees-C, pH 5.4).3. The pho-2A mutant of N. crassa secreted a more thermolabile acid phosphatase (t1/2 = 30 min at 50-degrees-C, pH 5.4) than the wild strain (t1/2 of at least 80 min at 50-degrees-C, pH 5.4).4. The pho-2A mutant of N. crassa synthesized a more thermolabile acid phosphatase (t1/2 = 37 min at 54-degrees-C, pH 5.4) than the wild strain in high Pi medium (t1/2 = 14 min al 54-degrees-C, pH 5.4).5. The pleiotropic nature of the pho-2 locus and its possible involvement in the mechanism of phosphatase secretion by N. crassa are proposed.

Kinetic properties of de-repressible alkaline phosphatase (EC 3.1.3.1) from Neurospora crassa

De-repressible alkaline phosphatase from N. crassa shows inhibition by PNP-P and a hyperbolic mixed-type inhibition by Pi. Both increasing concentrations of Pi and decreases in assay pH abolished inhibition by the substrate. Also, Pi promoted polymerization of the enzyme molecule, whose effect may account for the inhibitory behaviour shown by the enzyme in the presence of low Pi concentrations. © 1991 Rapid Communications of Oxford Ltd.

Exercise training decreases mitogen-activated protein kinase phosphatase-3 expression and suppresses hepatic gluconeogenesis in obese mice

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

CA15-3 and alkaline phosphatase as predictors for breast cancer recurrence: a combined analysis of seven International Breast Cancer Study Group trials

BACKGROUND: We evaluated the ability of CA15-3 and alkaline phosphatase (ALP) to predict breast cancer recurrence. PATIENTS AND METHODS: Data from seven International Breast Cancer Study Group trials were combined. The primary end point was relapse-free survival (RFS) (time from randomization to first breast cancer recurrence), and analyses included 3953 patients with one or more CA15-3 and ALP measurement during their RFS period. CA15-3 was considered abnormal if >30 U/ml or >50% higher than the first value recorded; ALP was recorded as normal, abnormal, or equivocal. Cox proportional hazards models with a time-varying indicator for abnormal CA15-3 and/or ALP were utilized. RESULTS: Overall, 784 patients (20%) had a recurrence, before which 274 (35%) had one or more abnormal CA15-3 and 35 (4%) had one or more abnormal ALP. Risk of recurrence increased by 30% for patients with abnormal CA15-3 [hazard ratio (HR) = 1.30; P = 0.0005], and by 4% for those with abnormal ALP (HR = 1.04; P = 0.82). Recurrence risk was greatest for patients with either (HR = 2.40; P < 0.0001) and with both (HR = 4.69; P < 0.0001) biomarkers abnormal. ALP better predicted liver recurrence. CONCLUSIONS: CA15-3 was better able to predict breast cancer recurrence than ALP, but use of both biomarkers together provided a better early indicator of recurrence. Whether routine use of these biomarkers improves overall survival remains an open question.