Historical Perspective

Because the activity of proteins controls all cellular functions, from metabolism to cellular architecture, signal responses, motility and replication, it could be argued that all pathologies stem from the dysfunction or inappropriate expression of proteins. This accepted, it follows that there must be a set of protein metrics to describe all forms of disease, a subset of which will be diagnostic, prognostic, or useful in monitoring disease status.

It is surprising, therefore, that at present diagnostic proteins for many common diseases, such as cancer, are almost nonexistent (Table 1). For cancer, this is in part due to lack of sensitivity but is largely through the heterogeneity of malignancies precluding the discovery of widely expressed markers. The estrogen receptor (ER) is currently the most reliable predictor of therapeutic response, based on hormone sensitivity in breast cancer (1). The epidermal growth factor receptor (EGFR) holds an equivalent but inverse prognostic value in conjunction with the ER status (2). Urokinase plasminogen activator (uPA), a serine protease causally involved in cancer invasion and metastasis, is also prognostic for poor outcome in breast cancer and other adenocarcinomas, such as gastric, colorectal, esophageal, renal, endometrial, and ovarian cancers (3). Carcinoembryonic antigen (CEA) is used in postoperative monitoring of colon carcinoma where preoperative levels are known (4). The increment of prostate-specific antigen (PSA) due to proliferation of cancerous prostatic epithelial cells is considered a useful marker of disease; indeed, the serine protease family of kallikreins (of which PSA is a member) have been implicated in the progression of endocrine tumors (5,6). Nevertheless, normal variation in PSA levels and its elevation in benign hypertrophy can confuse diagnosis, and more reliable markers are needed. More recently, prostate-specific membrane antigen (PSMA) and prostate stem cell antigen (PSCA) have been studied as clinical markers of prostate cancer, although neither of these proteins is entirely specific to prostate (7,8).

A major drawback in attempting to define a single marker for any particular cancer is the propensity for tumors to stop expressing proteins that have become deleterious. Thus the development of such markers as therapeutic targets, for example, c-erbB2 in a proportion of breast cancers, bcr/abl kinase in chronic myeloid leukemia (CML), is a logical step but the therapies themselves produce a selective pressure for mutations that no longer respond. These may take the form of nonexpressers (ER) or polymorphisms of the original marker (abl kinase); close monitoring of long-term treatments is required to detect resurgent disease. Both disease heterogeneity and mutability point to the need for a set of markers to be used in the monitoring and prognosis of long-term cancer. But the discovery of new markers is either serendipitous or through protracted effort, for example, extensive genotyping or subtractive analysis of antigens or cDNA collections. The ability to analyze proteins en masse in comparative disease versus normal studies would be a significant advantage in this respect.

To take another example, in Alzheimer's disease, diagnosis still remains within the domain of cognitive impairment analysis; there are no serum markers. In cerebrospinal fluid (CSF), elevated tau and decreased amyloid beta 42 have been correlated with marked deterioration in the disease and are potentially diagnostic (9). In a small subset of patients, genetic testing can suggest susceptibility to the disease; in particular, Apo E4 variant (10), mutations in presenilin-1 and -2 and beta-amyloid precursor polymorphisms (reviewed in Refs. 11,12), but there remains no proteomic prognostic markers for Alzheimer's disease.

The use of markers of toxicology is vital for clinical testing of new drugs. Measures of organ function in liver, kidney, and heart are routinely used in the toxicological evaluation of any new compound. Predictive markers of adverse events are highly desirable both in terms of patient safety and in reducing trial costs, and proteomic profiling enables comparison of new compounds with drugs that have established mechanisms of toxicity. Defining a toxicological profile may also enable targeting of the mechanism of toxicity in order to develop protective drugs. Serum and CSF are already collected in quantities containing sufficient protein, and these sources are proven to be amenable to proteomic analysis (13). Studies on serum are already established, and, because the aim would be to develop rapid readouts of early toxicity, a strong opportunity exists in this field.

The field of efficacy markers is also amenable to proteomics. The majority of drugs act by altering the activity of protein targets, for example, imatinib (Glivec) (14) acts by inhibiting the kinase activity of bcr-abl. This protein overphosphorylates proto-oncogene Crk-like protein (CRKL) in the disease state, and the efficacy of the inhibitor can be evaluated by measuring phosphorylation of this marker. Drug resistance is associated with mutation of the target kinase such that the inhibitor is no longer effective (15). Although relatively straightforward in the case of a circulating tumor, such as CML, in which blood extracts may be tested, monitoring of solid tumors is less reliable as, for instance, postoperative relapse is frequently at a remote and unpredictable site. The discovery of easily accessible, that is, circulating or excreted, protein markers altered by successful treatment of disease would provide a useful tool in molecular monitoring of drug efficacy. In the treatment of Gaucher disease the level of serum chitotriosidase is a significant marker of efficacy in both oral (miglustat) and enzyme replacement (Cerezyme) therapies (16,17), although measurement of this enzyme is also used to monitor beta-thalassemia (18). The use of single protein markers often proves inconclusive: Comprehensive comparative analyses of secreted or serum proteins can be used to define the efficacy of specific sets of markers or disease status, enabling more accurate monitoring of treatments.

Elucidation of the interactions and biochemical pathways mediated by proteins in vivo presents the next great challenge to biology and will have a revolutionizing effect on drug discovery. Genomics-driven target discovery has thrown up well-characterized genes, which await therapeutic drugs, for example, B-raf (19) and the breast cancer susceptibility genes, BrCal (20) and BrCa2 (21). Mutations in other genes, such as ras and tumor suppressor p53, induce constitutive activities affecting cell cycle and have well-documented roles in the process of carcinogenesis (22). The attractiveness of these proteins as targets is muted by their intractability in drug screens or chemical optimization; nevertheless, the cellular enzymatic and signaling pathways through which they effect abnormal replicative responses may prove amenable to modulatory therapy.

Heterozygosity, or polymorphic functional compromise of key enzymes or metabolite transporters in biochemical pathways, causes inherited errors of metabolism. In these cases the choice is either to replace the defective enzyme or to inhibit the buildup of metabolites caused by biochemical bottleneck by inhibiting another enzyme in the pathway. Thus, miglustat, an inhibitor of ceramide glucosyltransferase, has been shown to reduce the accumulation of glycosphingolipids caused by impairment of gluco-cerebrosidase activity in Gaucher disease (23). Proteomics offers a new means of discovering such enzymes either as targets or biomarkers. Metabolic and signaling proteins frequently form functional complexes on the surfaces of membranes or within organelles carrying out specific cellular activities: sensitive detection and microsequencing of these proteins is now achievable with the relatively small quantities of protein complexes enriched by affinity techniques. The discovery of well-characterized proteins in previously unknown settings can open the way to novel therapeutics.

Another aspect of genomic discovery has been the cataloging of gene families related to known drug targets. Indeed, such terminology as "druggable proteins" has been freely applied to homologous classes, such as G-protein coupled receptors (GPCRs), protein kinases, and ligand-activating proteases. Genomic (microarray) identification of these potential targets is rapid and relatively inexpensive; demonstrating protein expression and localization in disease tissues considerably less so. As a result, many proteins are suitable for pharmaceutical intervention, which have not yet been associated with disease. Proteomic analysis can provide a rapid route to validation of protein expression in the context of disease and therapeutic accessibility.

Therapeutic antibodies have recently risen in popularity within the pharmaceutical industry, which is attracted by their very high specificity, ease of synthesis, rapid lead selection, and (in the case of fully human antibodies) lack of toxicity relative to other

Table 1 Examples of Disease Biomarkers

Disease

Marker

Diagnostic

Prognostic

Disease monitoring

Prescription

Breast cancer

Her2neu

3

3

3

3

Breast cancer

ER

3

3

3

Breast cancer

EGFR

3

3

3

Breast cancer

BrCal, BrCa2

3

3

GI, renal, ovarian,

uPA

3

endometrial cancer

Colon cancer

CEA

3

Prostate cancer

PSA

3

CML

CRKL

3

CML

Bcr-abl

3

3

3

Alzheimer's

Tau, amyloid ¡542

3

3

Alzheimer's

Apo E4 variant

3

3

Gaucher's

Chitotriosidase

3

Gaucher's

Glucocerebrosidase

3

Abbreviations: ER, estrogen receptor; EGFR, epidermal growth factor receptor; CEA, carcinoembryonic antigen; PSA, prostate-specific antigen; CML, chronic myeloid leukemia; GI, gastrointestinal; CRKL, proto-oncogene Crk-like protein.

Abbreviations: ER, estrogen receptor; EGFR, epidermal growth factor receptor; CEA, carcinoembryonic antigen; PSA, prostate-specific antigen; CML, chronic myeloid leukemia; GI, gastrointestinal; CRKL, proto-oncogene Crk-like protein.

biomolecule therapies. Success in bringing a number of antibodies to the clinic for cancer and immune disorders has led to a recent switching of resources on the part of many pharma and biotech companies onto the discovery and development of these new medicines (24). Although antibodies only bind effectively to cell surface proteins, they are somewhat less constrained by the need to target a cellular function and can be used to "mop-up" soluble ligands or to direct immune-mediated destruction of cancer cells and other dysplastic tissues. Proteomic discovery has been used to generate datasets enriched for plasma-membrane proteins, using subcellular fractionation techniques and high throughput annotation, from which targets may be selected and rapidly validated for clinical relevance (25-27).

Unraveling Alzheimers Disease

Unraveling Alzheimers Disease

I leave absolutely nothing out! Everything that I learned about Alzheimer’s I share with you. This is the most comprehensive report on Alzheimer’s you will ever read. No stone is left unturned in this comprehensive report.

Get My Free Ebook


Post a comment