Editorial: Lessons From the Classic Scientific Literature

Early in my career, my graduate mentor, the late Howard S. Tager, told me that you know you are getting old when you begin framing your research talks in the context of scientific history. As I have aged (slowly), this certainly rings true for me now. Over the years, as I included more historical references in my scientific talks, I have often reflected on some important differences in the ways science is conducted today versus the past. One clear difference is that there is more multiple (>3) author papers published these days. The average number of authors per paper has risen from 1.5 in 1950 to more than 5 today (1). In our journal Molecular Endocrinology, the average number of authors per paper today is more than 6 (2). This increase in author numbers can be partially explained by the multitude of disparate techniques used in papers today that require the collaboration of many individuals. Our papers now are more complex, use many different types of experiments to test the hypotheses posited, and the conclusions are more nuanced with respect to molecular functions of proteins and nucleic acids. Nevertheless, reading some of the older scientific literature has taught me lessons that I think are valuable to any scientist today. So, let me reflect here upon one aspect of scientific history that has helped to shape the kind of scientist and diabetologist I have become.

“Our papers now are more complex, utilize many different types of experiments to test the hypotheses posited, and the conclusions are more nuanced…… Nevertheless, reading some of the older scientific literature has taught me lessons that I think are valuable to any scientist today”

Recently, at the Annual Diabetes Symposium in Indianapolis, I had the opportunity to review and present a short history of insulin research and how the scientists who studied this one small peptide hormone transformed our understanding of molecular endocrinology. In the nearly 100 years that have elapsed since the discovery of insulin by Banting and Best, the discoveries surrounding the insulin molecule have represented “firsts” or “among firsts:” insulin was the first protein to be sequenced, the first protein to be chemically synthesized, the first protein whose concentrations were determined by RIA, among the first proteins whose crystal structure was solved, and the list goes on. Two scientists whom I admired greatly for their pioneering work during this heydey of insulin-related discoveries, and whom I have cited in many of my research talks, died in the last year: Donald F. Steiner and Ronald E. Chance.

In 1967 at the University of Chicago, Donald Steiner discovered that the 2-chain insulin molecule is synthesized as a larger, single chain precursor that he termed “proinsulin” (3, 4). His studies used pulse-chase labeling experiments followed by size-exclusion chromatography and collectively showed the incorporation of label into a peptide larger than insulin during short periods of pulse and later converted to insulin during longer periods of chase. The following year in 1968 at the Lilly Research Laboratories in Indianapolis, where the identical precursor was previously identified as a “minor component” of crystalline porcine insulin preparations, Ronald Chance reported the amino acid sequence of porcine proinsulin (5). His studies required a series of carefully considered tryptic digestions, anion exchange column separations of peptides, and sequential Edman degradation to piece together the puzzle of C-peptide placement and sequence. Soon after these discoveries, it became evident that precursor, “proproteins” are common species predating more mature isoforms of a variety of peptide hormones and other signaling molecules. From a therapeutic perspective, the discovery of proinsulin and its sequence allowed for the commercial-scale production of active, recombinant human insulin (and their analogs) using a much more efficient “single chain” methodology. In reflecting upon these discoveries, I pondered how the approach to science has changed and what we can learn from reading these older papers. Certainly, the technology has changed, we are in the genomics era, and as such, the thought of running size exclusion chromatography columns or Edman degradation reactions to identify proteins and their sequences is unthinkable today (how many scientists today even know what these techniques are?). The thought of manually piecing together sequence data without the benefit of bioinformatics software tools seems almost barbaric, why risk making a mistake when a computer can do a better (and faster) job of it? So, yes, science is more technically advanced.

But, there are at least 3 good reasons why we should still read these old papers, and why we should especially encourage our trainees to read them. Perhaps the first reason is to think about how simpler experiments might test more precisely the hypotheses we posit. Today, if we were to posit that a single chain precursor of a polypeptide hormone exists, it would be tempting to initially test that hypothesis by an mRNA sequencing experiment (to identify an mRNA species larger than would be predicted by the polypeptide), or by a labeling mass spectrometry experiment. As state-of-the-art as these approaches might be, they are more expensive and more open to interpretation than performing a simple pulse-chase experiment, as Steiner did. Moreover, imagine how much more information you are getting than you actually need, or even whether the experiments are telling you what really wanted to know. The lesson here is that it is not a question of whether you could do an experiment but rather whether you should do the experiment. Second, and this is very important, by reading the older literature, we learn the basic principles that led to the techniques that we routinely use today. Chance had to devise a peptide separation strategy based on charge and size (using anion exchange chromatography); this technique is now a routine process for the kit-based isolation of many macromolecules, such plasmid DNA preparations on a large scale; however, how many of our trainees know that this is what they are actually doing when using plasmid isolation kits? How can you troubleshoot something you do not understand? The lesson here is, do not just follow instructions but know how and why your technique works. Finally, there is one other important reason to read the older literature: you may realize that someone discovered something before you did, only using older techniques. For example, in the islet literature, we often attribute the observation correlating human obesity (and insulin resistance) with islet hypertrophy to several histomorphometry studies from the past decade. However, this observation was actually originally made from autopsy studies performed in the early 1930s by Robertson F. Ogilvie (6), only he did it by cutting out and weighing photographs of islets from his histological specimens. Despite the advancement in technology, we get nearly the same result today that he got over 80 years ago. When we perform literature searches and ignore the much older papers, it has the unintended consequence of not giving credit to those whose ideas and findings preceded ours. In a reassuring trend, a group at Google, Inc recently evaluated trends in the citation of older scientific literature and noted that in 2013 the citations of papers in the health and medical sciences published more than 10 years previously occurred 30% more compared with 1990 (7). The authors concluded that with the availability of online databases, the recognition of older studies is increasing. The lesson here is that great, even exceptional, science was done more than 10 years ago (and sometimes much longer).

There are many ways we can learn and propagate these lessons from scientific history. As mentors and teachers, we can require that students read the classic literature as part of the coursework that we teach; we can ask that journal club presentations occasionally include an analysis of a classic paper from history; in the day-to-day lab setting, we can challenge students to learn the principles and origins of an experimental technique that they use routinely or that they ask others to do for them. It may sound like busy work, but it is not. There is now what seems to be an increasing trend to “farm out” many experiments to collaborators and commercial entities, and not just complicated ones like bioinformatics or deep sequencing but those that were once considered bread and butter, such as plasmid construction and recombinant protein purification. This was not the case in the days of Steiner and Chance. Although this trend serves to enhance collaboration, increase the pace of research, and make the science better, we must nevertheless guard against any tendency to “disown” the experimental techniques done by our collaborators and subcontractors.

In closing, my advice to the next generation of scientists is to embrace the science done by the generations that preceded them. I am not saying that everyone should frame their own science in the context of scientific history (you get to do that when you are sufficiently old) but rather learn why the scientists of previous generations did the experiments that they did, how they did those experiments, and why they came to their conclusions. It is more than an exercise, it is a challenge. And challenge is why we all pursued scientific careers.

Raghavendra G. Mirmira, MD, PhD