Evolutionary approaches to obesity and T2D are not new. There is a history of evolutionary hypotheses intended to do one or more of the above. The first major change in the thinking about diabetes that had a strong evolutionary flavor was brought about single-handedly by a person called James Neel in the early 1960s [1]. Most people at that time believed that diabetes was a genetic disorder much like other well-known genetic disorders such as sickle cell anemia or hemophilia. The belief was so deep rooted that there are papers claiming Mendelian nature of inheritance of type 2 diabetes [2–4], although we know now that T2D is far from being Mendelian. This amusing conclusion appears to arise out of some facets of human nature to which researchers are no exception. When there is a theory and you expect to see or find something based on it, frequently, you start seeing things although they may not actually be there. Also if you do not expect something or do not have a theory for it, you are very likely to miss seeing it, although it may very much be there (“We do not see what we sense. We see what we think we sense” [5]). I have experienced this frequently with undergraduate students. When they are told that the given sample is most likely to contain say staphylococci, every dust particle starts looking like Staphylococcus. Perhaps this is how the Mendelian inheritance of diabetes was “observed” by some researchers. Neel also appears to believe quite strongly that inheritance of diabetes was Mendelian. Still Neel appears to be the first person to be convinced that diabetes could not be a “genetic disorder.” Not that he doubted the “genetic” part of it, but he raised doubts about diabetes being a “disorder” of genetic origin. All monogenic (where a defective allele at a single gene locus can be clearly shown to be responsible for the disorder) heritable disorders always exist in any population in exceedingly small frequencies, typically one in several thousand. Diabetes, on the other hand, was much more common and notably increasing in frequency by this time. Realizing that no genetic disorder can be as ­frequent as diabetes is and moreover cannot increase in a population so rapidly, he thought that diabetes should not be treated as a genetic disorder. If the frequency of the diabetic allele is high, there must be a biological reason for it. A frequency of a gene can increase if it is useful in some way so that natural selection favors individuals having that allele. Therefore Neel thought that the genes responsible for diabetes must be adaptive under at least some set of conditions. Since the association of obesity with diabetes was well accepted, he equated the diabetic tendency to the tendency to accumulate fat and argued that a “thrifty” gene helped storage of fat under conditions of better availability of nutrients and allowed reutilization of it under starvation. Neel’s definition of thriftiness is “being exceptionally efficient in intake and/or utilization of food.” This gives two distinct meanings to thrift. Either a thrifty individual eats more and stores fat or has a less wasteful metabolism which allows greater storage of energy. Let us call the intake efficiency as type 1 thrift and utilization efficiency, i.e., lower rate of burning calories to store more, as type 2 thrift. This classification is extremely important to understand thrift, but most researchers have failed to clearly differentiate between the two, leading to much ambiguity. The two are mutually compatible, and Neel appears to imply both. But as we will see below, both do not apply equally to all arguments about thrift. Therefore, for every argument, there needs to be some clarity as to whether we are talking about type 1 or 2 thrift or both. In Neel’s view the thrifty gene was under positive selection pressure in ancestral life when seasonal and climatic conditions resulted into fluctuating food availability often called “feast and famine.” The assumption is that human ancestors frequently faced periods of starvation interspersed with periods of food abundance. A genetic tendency to store fat when food was abundant was extremely adaptive under feast and famine conditions and therefore enjoyed a selective advantage. In modern life food is always abundant, something that can be called a “feast and feast” condition, and under such conditions, the same genetically determined tendency leads to obesity. Both types 1 and 2 thrift are compatible with this argument. According to this theory, obesity is the apparent cause of diabetes, but the diabetic tendency is the real cause of obesity. Neel appears to have believed that diabetic tendency leads to obesity. In his own words, “The overweight individual of 40 or 50 with mild diabetes is not so much diabetic because he is obese, as he is obese because he is of a particular (diabetic) genotype.” What Neel refers as mild diabetes is type 2 diabetes in the current nomenclature. The parenthesis for the word “diabetic” is original. Neel’s argument was based on a number of patterns well known by then including that diabetic mothers (also diabetic fathers!!) have larger babies, that hyperinsulinemia precedes adult-onset diabetes, that insulin has lipogenic action, and also that frequently there is an early history of transient hypoglycemia much before diabetes sets in. Neel thought that hyperinsulinemia is the thrifty tendency. High levels of insulin induce lipogenesis and fat accumulation since it facilitates entry of glucose into fat cells and its conversion to fat. One of his relied evidence was an early hypoglycemic phase in the history of T2D. This hypoglycemia perhaps represented active lipogenic metabolism under the influence of insulin. He thought that at later stages, the high levels of insulin were compensated by “anti-insulin” activity (which could be insulin resistance in today’s terms), and eventually, somehow, anti-insulin activity became stronger than insulin action to cause diabetes.

Although I am going to argue against some of Neel’s concepts, I must admit first that this paper is one of the greatest papers ever written on diabetes so far. It is full of prophecy and foresight. Neel does not hesitate to make bold speculations whenever he is confident about the logic behind them. It is doubtful whether in today’s science publishing trend, a paper as speculative as this one would have ever got published. Speculations are generally unwelcome in today’s science, and speculative papers are most likely to be rejected by the current editorial standards. Perhaps things were different in the early 1960s or Neel was extremely lucky. Neel’s speculations had a far-reaching vision. Interestingly when Neel wrote this paper, the concept of insulin resistance was not yet established. But Neel appears to have perceived it speculatively and calls it “anti-insulin” activity. Neel also had a very clean insight into a number of issues with which the followers of the thrifty gene hypothesis later messed up completely. Neel seems to be clear in his mind that hyperinsulinemia comes first and anti-insulin activity appears consequently. He also has rejected the idea of β cell “exhaustion” leading to diabetes.

Neel’s concept of thrift soon became almost an axiom, although no such “thrifty” gene or set of genes have been convincingly demonstrated any time. Later the observation that individuals born small for gestational age had a greater probability to become obese and type 2 diabetic in later life led to the concept of fetal programming [6–10]. This hypothesis states that if a fetus faces inadequate nutrition in intrauterine life, the body is programmed to be “thrifty” as an adaptation. Although a little later I am going to contest this logic as well, let us first appreciate the great vision of the researchers who detected this pattern for the first time. It is not easy even to imagine that the roots of a disorder that becomes obvious in one’s 40s and 50s could have its origins in intrauterine life. But a group of British visionaries discerned such a strange relationship with insightful analysis of data. The concept originated in data on differential death rates in different parts of Britain. Death rates in the newborn in the early decades of 1900s were highest in some of the northern industrial towns and poorer rural areas in the north and west. This geographical pattern closely resembled that in deaths due to coronary heart disease decades later. This led to the suspicion that intrauterine growth retardation had something to do with heart disease in adults. Epidemiological studies were launched based on the simple strategy of examining men and women in middle and late life whose body measurements at birth were recorded. Sixteen thousand men and women born in Hertfordshire during 1911–1930 were traced, and the effort revealed that death rate due to coronary heart disease in the lowest birth weight class was double than in the uppermost class [8]. Later, similar trends were found across most parts of the world and different ethnic groups, showing the robustness of the relationship between fetal growth conditions and adulthood disorders including obesity, type 2 diabetes, hypertension, and coronary heart disease [11–17]. There are two possible components of the proposed thrifty adaptation in response to reduced fetal growth. One relates to an immediate gain in terms of survival during fetal and early infant life. The other is a predictive adaptive response in anticipation of starvation in later life [18]. This distinction is important in understanding the evolution of fetal programming as we will see soon.

Both the concepts of thrifty gene and thrifty phenotype by fetal programming have recently faced serious criticism on several grounds [19–24]. Some of the critics think that the thriftiness concept is flawed and needs to be abandoned [20, 21]. Others have attempted to refine the concept of thrift so as to resolve some of the flaws and paradoxes pointed out by the critics [25–28]. Let us see the classical concept of thrift and its criticism first before considering the refined versions of thrift. Baig et al. [29] integrated the critical arguments challenging the classical thrifty gene and thrifty fetal programming hypotheses using a mathematical model. The model considered three hypothetical genotypes, namely, a non-thrifty wild type having no mechanism for thriftiness, a thrifty genotype which is genetically programmed for thriftiness, and a genotype with a capacity for fetal programming for thriftiness. Taking a year as a natural time unit of seasonality, a simple dichotomy of years was assumed in the model, those with adequate food supply (feast) and those with inadequate food supply (famine). Famines were assumed to occur randomly with some probability. The fitness of an individual with non-thrifty genotype in feast conditions was assumed to be greater than that of an individual with thrifty genotype because of a cost associated with thrift. Obesity and insulin resistance are known to be associated with reduced fecundity [30–33] justifying the cost of thriftiness in feast conditions. Fitness of an individual with non-thrifty genotype in famine conditions was assumed to be less than that of individuals with thrifty genotype. It is simple to visualize from this set of assumptions that at low probability of famine, the non-thrifty genotype will have an advantage, and at high frequency, the thrifty one will have an advantage. It also follows that at no condition, non-thrifty and thrifty genotype would coexist in a stable polymorphism. This is simple to see intuitively as well as to show with simple mathematics as done by the Baig et al. model [29].

Calculation of the fitness of the thrifty programmer was a little more complex. Assuming no correlation between birth and lifetime conditions, the total fitness was written as the sum of all years with the assumption that in the birth year the phenotype was best suited for given conditions. For the rest of the lifetime, the fitness fluctuated according to randomly fluctuating environmental conditions. Analytical solutions to the model showed that at low probabilities of famine, a non-thrifty gene has a selective advantage, and at high probability, thrifty gene would get selected (Fig. 4.1) leaving a very narrow area of advantage for fetal programming. The area was narrower for long-lived species, whereas for short-lived ones, the birth year advantage was large enough as compared to lifetime, and therefore, fetal programming had a much larger width of advantage (Fig. 4.2). Two interesting conclusions of the model are that a thrifty gene can evolve but cannot have a stable polymorphic state and that fetal programming for thrift is most unlikely to evolve by feast and famine conditions in long-lived species [29]. There are a number of other issues about the concept of thrift that need to be addressed.