There have been extraordinary individuals active in patient care, clinical research, professional and patient organizations, and pharmaceutical science who devoted their careers to constantly focusing on and campaigning for “what is best for the patient.” Dr. John A. Galloway has been that leader, whose clinical curiosity, insights, dedication, and commitment to others brought insulin therapy from the impurified, low-potency animal insulins to the modern insulins of today.
Born in Omaha, Nebraska, in 1928, Galloway moved to Washington, DC, and graduated from St. Albans School in 1946 and then from the University of Pennsylvania in Philadelphia in 1950 with a B.A. in English and Pre-Med. After 2 years in the U.S. Army and serving in combat in Korea, 1st Lt. Galloway was released to start medical school at the University of Nebraska College of Medicine in Omaha, where he gained his MD in 1956. Upon completion of his internship, he began his residency in internal medicine at Temple University Hospital in Philadelphia, serving as Chief Medical Resident from 1959–1960. This was followed by a Fellowship in Diabetes, Endocrinology and Nuclear Medicine with Charles R. Shuman, MD, at Temple, where Galloway and Shuman published the first report in the medical literature on the use of computers to analyze large volumes of clinical data in a study of surgery in patients with diabetes (1).
After a report in the Journal of Clinical Investigation by Solomon A. Berson, MD, and Rosalyn S. Yalow, PhD, a physicist, on a method they had developed for measuring insulin in the blood using immunologic techniques (the radioimmunoassay or RIA) (2), Shuman arranged for Galloway to attend the first workshop that Berson and Yalow gave on how to perform their assay. Also selected to attend were Lisa G. Heding from Novo Laboratory (now Novo Nordisk) in Denmark, and Roger Unger, MD. Heding went on to transform the Berson and Yalow assay into one that was fast and efficient and therefore useable in most dedicated research laboratories. Unger later achieved the exceedingly difficult adaptation of the RIA for the measurement of serum glucagon, leading to a distinguished career in the roles of glucagon in diabetes. Another individual key to the development of RIA technology was Mary A. Root, PhD, from the Lilly Research Laboratories. Root’s work on the RIA focused on the preparation and purity of the insulins for use both in inducing anti-insulin antibodies in the appropriate animal models and the preparation of the radiolabeled tracers to be used in the assays. Most investigators in the early days of insulin and glucagon RIAs (including this researcher) relied on Root for supplies of pure hormones to conduct their research.
An Interest in Insulin Immunology
Prior to the Berson and Yalow RIA, the only assays for insulin in the blood were biologic, usually measuring production of C14 CO2 from C14 glucose in rat hemidiaphragms or epididymal fat pads. Those techniques were extremely cumbersome and required great technical skill to perform. As a result, the numbers of samples that could be assayed were extremely small.
Berson and Yalow developed their assay from observations during their examination of blood from patients with immunologic insulin resistance, who required very large doses of insulin. At that time, insulin was made from animal pancreases, usually beef alone or beef and pork, and it commonly provoked formation of antibodies when used therapeutically. They observed that when radiolabeled insulin was mixed with the serum from a patient with immunologic insulin resistance and the combination was placed on filter paper and analyzed by an electrophoretic technique, two peaks were discernible. One occurred at the origin, or point of application, on the filter paper. They identified this as “free insulin.” The other peak occurred a few centimeters from the point of origin. They identified this as antibody-bound or “bound insulin.” They simply used scissors to cut out the two peaks, weighed the two filter paper segments on a scale, and came up with a bound/free ratio. Utilizing the principle of differing patterns of migration of free and bound insulin, they immunized guinea pigs to generate anti-insulin antibodies. By adding known amounts of insulin to their system they were able to generate a standard curve to assess the amount of insulin in serum samples (2).
While their assay was cumbersome, requiring storage of the impregnated filter-paper strips for 24 h at 40°F and meticulous attention to accurate cutting, they produced the first proof of the presence of hyperinsulinemia and demonstrated insulin resistance was a key feature of both obesity and type 2 diabetes (3).
This was the nidus that stimulated Galloway’s lasting interest in insulin immunology, leading to his long-term collaboration with S. Edwin Fineberg, MD, and Naomi S. Fineberg, PhD, of Indiana University.
Later in his career Galloway would note that Christian Binder, MD, of Nordisk Laboratory (now Novo Nordisk), reported on the pharmacology of insulin using the RIA to measure serum hormone levels after the injection of doses of therapeutic insulin. It was Binder’s inspiration and Root’s assay that made it possible for Galloway to study the various new forms of insulin being developed by the Lilly Research Laboratories.
Improving Patient Care was Always Galloway’s Top Priority
After his fellowship and following a brief private practice stint, Galloway was recruited by William R. Kirtley, MD, in April of 1962 to join Eli Lilly and Co. in Indianapolis to run clinical programs for diabetes. Galloway became Director of Clinical Diabetes at the Lilly Clinic (Lilly Laboratories for Clinical Research), then located at Wishard Memorial Hospital on the campus of the Indiana University School of Medicine, where he was appointed Associate Professor. Galloway’s interest in improving the lives of patients resulted in the clinic accepting patients with unusual and/or extremely difficult-to-control diabetes for evaluation.
After evaluating a number of elderly patients with poor control but few complications, Galloway suspected that parental longevity might be a protective factor against the physiologic damage of diabetes. In 1974, Harry A. Cochran Jr., MD, Medical Director of the Lincoln National Life Insurance Co. in Fort Wayne, Indiana; Alexander Marble, MD, at the Joslin Clinic in Boston; and Galloway sent questionnaires to 99 individuals who had been on insulin for 50 years or more. Using his medical statistical expertise, Cochran analyzed the returned report forms going back two to three generations. The group’s conclusion: recent ancestral longevity seemed to protect against the ravages of diabetes and was the single most significant factor in predicting longevity in people with diabetes (4).
Another finding was in a 38-year-old African American woman who was referred to Galloway from Vanderbilt University for treatment of insulin resistance. She required 3,000–5,000 units a day and oral sodium bicarbonate to prevent acidosis. Not finding evidence of increased serum antibodies to insulin, Galloway and fellow Jaime A. Davidson, MD, reasoned that she had a defect in her insulin receptor system and sent her to the National Institutes of Health (NIH) Diabetes Research Center in Bethesda, Maryland, placing her under the care of Philip Gorden, MD. The NIH evaluation disclosed antibodies that impeded the interaction of insulin with its receptors. Studies there on several patients resulted in the first article on insulin receptor defects by Kahn et al. (5).
Galloway’s greatest contributions to improving the care of patients with diabetes were in the area of development and formulation of commercial insulins, more specifically, their pharmacokinetics and pharmacodynamics, purity and immunogenicity, and source, supply, and availability.
The insulin concentration in the first vials of commercial insulins in 1923 was 4–5 units/mL and later 10 units/mL. The first “regular” insulins were tan to brown in color, acidic, and often caused significant pain and allergic reactions. Over the years, the concentration and purity of commercially available insulin increased to 40 and 80 units/mL. At the same time, in order to reduce the frequency of four to five daily injections, various substances were added to insulin formulations to extend the action, thus reducing the number of injections required by patients. Protamine and/or zinc were used as extender agents, resulting in the creation of protamine zinc insulin in 1938, isophane or NPH (neutral protamine Hagedorn) insulin in the late 1940s, and the zinc insulins (Lente, semilente, lente, and ultralente developed by Hollas-Mueller in the 1950s). An insulin manufacturer desiring to carry a full line of insulins would have to market at least 12 different forms of insulin. By itself, the number of insulin products was not a problem.
However, there was a problem with syringes and the hazard of patients confusing insulin strength, resulting in a significant incidence of unwarranted hypo- and hyperglycemia. There were syringes calibrated to 40 units/mL on one side of the barrel and to 80 units/mL on the other side. Galloway and colleagues proposed converting the insulin line to a single strength of 100 units/mL (or U-l00). The first step was to gain the assurance from syringe manufacturers that a syringe that accurately delivered a given dose of insulin containing 100 units/mL could be made.
Galloway started this process by meeting with Bruce Griffith at Becton, Dickinson, and Co. Soon the six U.S. syringe manufacturers confirmed that they had the technology to make syringes of a sufficiently small bore to deliver small doses of U-100 insulin accurately. After gaining support from the American Diabetes Association (ADA), proposals were made to the U.S. Food and Drug Administration (FDA) and the Insulin Committee in Toronto. Shortly thereafter the Danish insulin manufacturers (Novo and Nordisk Laboratories) along with the British Diabetes Association were in agreement. It was the Danish insulin companies and the British Diabetes Association who convinced the other European diabetes groups (including the European Association for the Study of Diabetes) and individual country diabetes organizations to accept the U-100 insulin (6).
When the U-100 proposal was discussed at the International Diabetes Federation meeting in Brussels in 1973, pediatric diabetology specialists expressed concerns about the practicality of accurately delivering small doses of U-100 insulin. The insulin manufacturers agreed to provide gratis diluting fluids, mixing vials, and syringes for such patients. The U-100 insulin conversion program went smoothly, and by the mid-1980s U-40 and U-80 insulins were withdrawn from the market.
Galloway—Lead Physician in the Clinical Development of Human Insulin
At a symposium held in Indianapolis in 1972, on the occasion of the 50th anniversary of the discovery of insulin, concerns were expressed about the adequacy of the supply of insulin from animal pancreases to meet the long-term needs of a growing population of patients with diabetes and increased insulin use. At the same time the per capita consumption of beef and pork was decreasing worldwide. Novo and Lilly were competing in the U.S. market for animal pancreases. The interest and expertise of Irving S. Johnson, PhD, at the Lilly Research Laboratories in Indianapolis, in islet cell cultures that produced insulin proved a significant first step toward finding a new source of insulin. It was Johnson who first recognized and acted upon the possibility that the discovery by James Watson and Francis Crick (who had broken the genetic code) would soon be followed by DNA sequencing and gene insertion into various organisms to produce peptides including drugs such as insulin. He recognized that such activities would require collaboration with outside investigators. In 1978, David Goeddel, PhD, and colleagues at Genentech produced the first human insulin using A and B insulin peptide chains expressed in Escherichia coli from genes that had been chemically synthesized by Itakura and colleagues at the University of California, San Francisco (7). Collaborations were established with these institutions, and patent agreements were signed that would permit Lilly to scale up the process of human insulin production from 20 ng in California to 5,000 kg of insulin in Indianapolis in 4 years at an upfront cost of over $300 million for fermentation and purification facilities.
Johnson’s staff consisted of scientists and engineers from every discipline. Galloway served as the lead physician, under the direction of John H. Marsden, MD, in the clinical development of human insulin (later branded as Humulin). The clinicians gained tremendous knowledge from Ronald E. Chance, PhD, and Bruce H. Frank, PhD, pioneers in insulin chemistry and protein formulation.
Concurrent with the scientific development, additional enormous tasks were undertaken by Johnson and Galloway to obtain approval by various scientific and regulatory agencies in the U.S. and Europe. These included the National Academy of Science, the NIH, the FDA, and regulatory agencies in Europe. An initial concern was containment of the fermentation process. To address this concern, it was agreed that all experiments would be conducted in a facility with the highest level of biological containment. There was the specific question, “Would E. coli organisms containing insulin cause hypoglycemia if they reached the gastrointestinal tract?” Fortunately, in 1971 Galloway and colleagues had performed experiments with orally administered insulin (5 and 10 units/kg) in normal subjects and demonstrated no effect except under the most extreme conditions. These conditions included atropine blockade of gastric acid secretion and having the volunteers lie on the left side, where the insulin solution could pool in the greater curvature of their stomachs (8).
The clinical testing of human insulin of rDNA origin (HIrDNA) was facilitated by two facts. First, the protein entity, human insulin, had been fully characterized. Differing from porcine insulin in only one amino acid at the B30 position (threonine instead of alanine) human insulin could be expected to act in humans much as pork insulin. Second, they had just completed extensive “new patient” and “crossover” studies comparing pork insulin of differing levels of purity because “mono-component” pork insulin was evolving to become the new standard of purity. Thus, the clinical investigation centers that would be evaluating human insulin were already up and running. Following phase I comparisons of pork and human insulin, first by Prof. Harry Keen at Guy’s Hospital and later at the Lilly Clinic by Daniel C. Howey, MD, Galloway’s group was prepared to embark on multicenter studies.
The first dose of human insulin of rDNA origin was administered by Dr. Richard A. Guthrie in Wichita, Kansas, on 16 December 1980. The 6-month crossover studies were completed within a year, and the data was submitted to the FDA. The FDA approved human insulin of rDNA origin in mid-1982 after only a 6-month review.
Many factors contributed to the fast review by the FDA. First was the previously mentioned concern about potential future animal insulin supply shortages. Second, the project benefitted greatly from the foresight and effectiveness of the FDA physicians and scientists like Henry I. Miller, MD, an expert on recombinant DNA technology at the NIH who moved to the FDA, and Solomon Sobel, MD, Director of the Division of Metabolism and Endocrinology Products at the FDA, who in anticipation of the new requirements had hired the expertise needed to review rDNA submissions.
Galloway believes a third force driving the rapid progress and approval of HIrDNA that was never discussed publicly was the possibility of the vulnerability of the animal insulin supply to contamination by Creutzfeldt-Jakob–like prion disease, such as bovine spongiform encephalopathy (mad cow disease) (J.A. Galloway, personal communication). The Division of Metabolism and Endocrinology Products of the FDA was clearly aware of this possibility, having already seen reports of such contamination of human pituitary glands that had been harvested for human growth hormone. It is not difficult to envision the enormity of the disaster if patients ever had any reason to doubt the safety of their animal-sourced insulin. According to Galloway, “Concern about the long-term adequacy of supplies of animal pancreas notwithstanding, I always thought the threat of Creutzfeldt-Jakob disease to be the occult driving force in the FDA’s rapid processing of the HIrDNA application” (J.A. Galloway, personal communication).
Galloway along with Davidson and Howey had a long history of performing studies evaluating the pharmacokinetic and glucodynamic effects of varying the species, concentration, and absorption modifying components of both commercially available and investigational insulins. With human insulin available, he explored the question of whether human insulin would be nonantigenic, as had been postulated for many years. Having seen antigenicity in the crossover registration trials, he designed a study with James Anderson, MD, then at San Antonio, and Ed Fineberg, MD, at Indiana University to randomize patients with new-onset type 1 diabetes to either multiple injections (or pump therapy) of regular human insulin only or combined use of regular and NPH human insulins. The regular human insulin–only patients demonstrated no antibody response during the 6-month trial, while those receiving doses of NPH human insulin had a classic (but low compared with animal NPH) antibody response. This was the first in a series of studies that would later confirm that the time of insulin residence in the subcutaneous tissue is related to the antigenicity response (9).
Not too long after the approval of HIrDNA made from combining insulin A and B chains, it was decided a more efficient method of making human insulin was to produce human proinsulin by fermentation and then enzymatically cleave the connecting peptide, resulting in producing both human insulin and human C-peptide (HCP).
This innovation made it possible for Galloway to study the potential of human proinsulin (HPI) to be used as a therapeutic tool in diabetes. HPI was found, after subcutaneous injection into animals and humans, to be slower acting than neutral regular human insulin and suggestive of a time-action profile similar to that of human NPH. With a molecular weight of about 9,300 KD and a potency of only 3 units/mg (human insulin is about 28 units/mg), on a molar basis HPI was substantially less active than human insulin. Nonetheless, the time action of HPI was attractive. Galloway began phase II studies in about 120 patients who had never received insulin and were randomized to receive either NPH human insulin or HPI. Unfortunately, no clinical benefit was seen, and the clinical evaluation of HPI was discontinued (10,11).
Galloway realized two possible uses for HCP. The production of human insulin from HPI resulted in the annual production at that time of ∼500 kg of HCP. First, at the behest of Arthur H. Rubenstein, MBBCh, and Kenneth S. Polonsky, MD, at the University of Chicago, Galloway supplied HCP for these investigators to develop a method of assessing endogenous insulin secretion by measuring endogenous C-peptide. The problem they were addressing was that the metabolism of endogenous C-peptide, just like insulin, was altered by the liver and the kidney. They reasoned that these latter two influences could be quantified by assessing the fate of exogenously administered C-peptide. From this information they could ascertain the amount and time course of endogenous insulin secretion, essentially simulating measurement of insulin as it is secreted into the portal vein. Their ingenious studies, which were enhanced by complex mathematical models, produced spectacular results in confirming some and disputing other findings about endogenous insulin secretion (12).
The other use of HCP was based on the possibility that the microvascular complications of type 1 diabetes were due to C-peptide deficiency. (Although commercial animal insulins before the mid-1970s were “contaminated” by C-peptide and other proinsulin-like derivatives or substances, the actual amount of C-peptide was quite low.) The question was whether there could be a benefit from adding HCP back to commercial insulin, thereby making therapeutic insulin more like normally secreted endogenous insulin. John Wahren, MD, PhD, at the Karolinska Institute in Stockholm, produced sufficient short-term data in animals to suggest this was possible. Neil White, MD, and the late Julio Santiago, MD, at Children’s Hospital in St. Louis, undertook a double-blind crossover study where patients with type 1 diabetes were treated with standard HIrDNA for 3 months and then switched to HIrDNA to which HCP was added in amounts equimolar to the insulin in the vials. The main end points were renal effects. They observed no effect of adding HCP. One may criticize this study, as Wahren did, for being too short in duration (J.A. Galloway, personal communication). Nonetheless, further evaluation of HCP as an additive to insulin was discontinued.
As the most senior Lilly physician, with the executive title of Research Fellow, Galloway provided critical direction for the development of Lilly’s human insulin analog programs (13). Galloway and colleagues addressed the fact that injected insulins failed to simulate normal insulin secretion. They knew from studies in normal volunteers that within a few minutes after ingestion of a meal, serum endogenous insulin secretions rose promptly to 80–100 μU/mL and fell toward baseline within 30–45 min. On the other hand, following the injection of neutral regular human or pork insulin, peak serum concentrations did not occur until 45–60 min and persisted for up to 8 h. They knew that the delayed effects of regular insulin were related to the tendency of the hormone to self-associate to a larger molecule in the injection site, thereby reducing its rate of absorption. Based on insulin-like growth factor work done at Lilly, Ronald E. Chance, PhD, Bruce H. Frank, PhD, James E. Shields, PhD, and Richard D. Di Marchi, PhD, in collaboration, synthesized the human insulin analog lispro by reversing the position of the proline and lysine amino acids at positions 28 and 29 of the B-chain, resulting in an insulin that did not self-associate. Subsequent global clinical work resulted in the introduction of insulin lispro, the world’s first approved rapid-acting human insulin analog.
Galloway is an author or coauthor on over 80 scientific articles and book chapters, an author of two books, and a collaborator on multiple patents. Board-certified by the American Board of Internal Medicine, he is Professor of Medicine, Emeritus, at Indiana University School of Medicine. He cofounded the Indiana Diabetes Association and served two terms on the Board of Directors of the ADA. He has been an active member of the International Diabetes Federation and the European Association for the Study of Diabetes and is a Fellow of the American College of Physicians.
He was awarded the Josiah Kirby Lilly, Sr. Distinguished Service Award from the Indiana Affiliate of the ADA in 1998. Recognized as an international ambassador for diabetes, he has been honored with a named lectureship from the International Diabetes Federation. For 30 years Galloway was the face and voice of Lilly Diabetes.
Even after retirement in 1993, Galloway remained active in diabetes advocacy, and in his “spare time” served his community as Co-Chair of Special Gifts for United Way in Cambridge, Massachusetts, and as a volunteer physician at the Senior Friendship Clinic, a free care facility in Naples, Florida.
On the occasion of Galloway’s 90th birthday, his colleagues, fellows, students, friends, and the community of individuals affected by diabetes would like to thank him and congratulate him on a job well done.