Book Review: Metabolic Profiling: Disease and Xenobiotics (Volume 21 of Issues in Toxicology)

For a book aimed at a toxicology audience, the book title is almost (and unfortunately) misleading or at least ambiguous. I suspect that many toxicologists upon reading the title, “Metabolic Profiling,” would envision a treatise on pathways of xenobiotic metabolism. This, however, is a book on metabolomics, that is, study of the changes in the intermediary metabolism (that may be caused by disease, toxicity therapeutic drug action, changes in diet, etc) and presented as changes in intermediary metabolites such as amino acids, sterols, simple sugars, volatiles, free fatty acids, and so on in the urine, plasma, or other bodily fluids. Fortunately, the ambiguousness of the title did not make for a poor book, and it is in fact a good book, but, with caveats.

There is some debate if not confusion in the literature on the subtle differences, if any, between “metabonomics” versus “metabolomics” (see Nicholson et al ¹ or Robertson, 2005). Depending on the author, 1 discipline has been described as a subset of the other. From a toxicologist’s point of view, however, these terms can be, and often are, used interchangeably. Since the term used in this book is metabolomics, this is the term used in this review. The utility of metabolomic profiling in the field of toxicology to provide mechanistic understanding or biomarkers for the toxicity of drugs has been discussed for at least a decade (see review by Robertson ² ).

From the preface, the editor states “This book represents the culmination of at least several years relatively intensive work and provides an in depth and sometimes highly critical review of research investigations performed in the metabolomics research area …” Basically, the focus of this book is on the appropriate experiential design and the application of appropriate statistical analysis of the metabolomics-based experimentation. Consider that a urine sample can contain hundreds of different intermediary metabolites and that these have to be sorted in the context of many variables (age, gender, time of day, age, nutritional status, presence of concurrent disease, etc). The amount of data that can be collected on relatively few samples is prodigious. It is the book editor’s (who is also one of the main authors of the book) contention that, heretofore, the metabolomics studies that he has read in the literature have not been well designed, and the data were poorly analyzed. He has seen a statistical error and is out to correct it. Thus, the first 5 chapters provided an in-depth coverage of experimental design and statistical analyses. Terms such as principal component analysis, multivariate analysis, Bonferroni correction, bootstrap aggregation, power analysis, and similar such terms abound. If you are not active in the field of metabolomics or statistics, have a book on statistics in hand and be prepared to spend your time reviewing these concepts as you study this book.

In the first 5 chapters, there is little mention of the underlying chemical analytical techniques used to generate those data. The authors have assumed that the readers have a good enough understanding of those methods to grasp the analytical approaches. Please refer to any 1 of the 3 basic reviews on the analytical techniques if you need a refresher. ^{1 –3} The actual analytical techniques are not discussed until Chapter 6, and the presentations focus on the more recent (and at times esoteric) advances in the field. Additionally, one would need a good understanding of the first 5 chapters to grasp the data analysis and the presentation used in conjunction with these analytical techniques. For example, Chapter 6 reviews the diverse applications of the different mass spectral (MS) techniques, reviewing the advantages and disadvantages of the technology depending on the biological samples in question. It is, in fact, a very good review of this particular subject. Chapter 7 discusses the use of MS (mass spectrometry), FT-IR (Fourier transformed infrared spectroscopy), and ¹H-NMR (Proton nuclear magnetic resonance) of biological fluids and the appropriate data analyses necessary to present such data. For example, ¹H-NMR used in conjunction with principal component analysis and partial least square discriminatory analysis has been applied to the characterization of brain metastases to develop potential diagnostic tools. Chapter 8 continues the theme of advanced analytical technologies, introducing and reviewing the advantages of group-specific internal standard technology for mass spectrometry-based metabolomic profiling. This involves the addition of isotopically labeled internal standards such that MS can now be used for quantitating dynamic changes in specific metabolites. For example, the use of ¹⁸O and ³¹P labeling in conjunction with MS for “phosphometabolomic” “finger printing” and metabolic monitoring is presented in depth.

Chapter 11 would perhaps be the most germane to a toxicologist. Detailed descriptions of the application of metabolomics techniques to drug-induced liver disease (DILI) are presented and discussed. It includes a very good review of the molecular underpinnings of most forms of DILI and then goes on to discuss in detail the advantages of differing metabolomics approaches to investigate and differentiate these pathways. For example, the oxidative stress associated with acetaminophen I results in increases in urinary creatinine and serum ophthalmate but a decrease in urinary S-adenosylmethioine. This chapter makes a compelling case for the use of such finger print-type shifts in endogenous intermediary metabolites as biomarkers of liver toxicity.

From a mechanical point of view, the book’s major weakness is the abundance of run-on sentences. In their haste and excitement on the subject material, the authors frequently tried to pack too much into 1 sentence, which at times ran on for lines at a time, and made already complex subject material even more difficult to follow. Here is an example from page 69.

Fundamentally, Moseley (2013) very recently described and reviewed the involvement and employment of error analysis in MV metabolomics explorations as an improvement in overall experimental design (which are generally poorly accepted or implemented in many published investigations), and hence, the prior consideration of appropriate statistical methods for their analysis (which should, of course, include validation and cross-validation models via mutation techniques, where relevant), essential quality control monitoring of the laboratory experiments performed and finally determination of our confidence (and hence potential uncertainties) in the results required

All in all, this is not a book for the faint of heart. It is not a survey or a review book but an in-depth critique and presentation of complex material. Reading this book may take some time, and you may need to have a reference material available while you do so. I am not ashamed to say that I did. On the other hand, if you are willing to take the time necessary to read, study, and understand this book, you will come away with an in-depth understanding of the statistical treatment of large data sets in general and metabolomic data in particular.