“Patients Can Have as Many Gene Variants as They Damn Well Please”: Why Contemporary Genetics Presents Us Daily with a Version of Hickam's Dictum

“The constant characters which appear in the several varieties of a group of plants may be obtained in all the associations which are possible according to the laws of combinations” (1). It may appear that this statement by Mendel, said more than 100 yr ago, flies in the face of Occam's “plurality must not be posited without necessity,” stated in 1318 by William of Ockham and essentially deified in modern medicine after its endorsement by Sir William Osler (2, 3). It has been suggested that in contemporary medicine, “Hickam often triumphs”(3). Dr. Hickam, the anti-Occam of medicine, is said to have stated: “patients can have as many diseases as they damn well please” (2). As is often the case in life, truth lies in between: a simple explanation for all matters of an illness (or a “phenotype”) is always attractive, but biology and pathophysiology are complex and depend on many variables that may not be ignored. This is particularly true when genome-sequencing studies indicate that each one of us carries as many as 100 loss-of-function mutations with more than 20 genes completely inactivated (4). On the other hand, in systems biology, pathways converge, and so at the end, many genetic variants are found to have previously unsuspected functional redundancies (4, 5) that essentially fit the Occam's principle: a patient can have as many genetic variants “as he damn well pleases” within a singular signaling pathway leading to multiple and/or graded phenotypes.

Which brings us to a classic description of a particular phenotype—that of primary adrenocortical insufficiency (pAI) by Dr. Addison in 1855: “The leading and characteristic features of the morbid state…are anemia, general languor and debility, remarkable feebleness of the heart's action, irritability of the stomach, and a peculiar change of color in the skin, occurring in connection with a diseased condition of the ‘suprarenal’ capsules…. This discolouration pervades the whole surface of the body, but is commonly most strongly manifested on the face, neck, superior extremities, penis, and scrotum, and in the flexures of the axillae and around the navel” (1, 6). As astute as this description is (everybody who sees patients with pAI can relate to this particular pattern and distribution of pigmentation), it has been recognized for a number of years now that patients presenting with pAI “need not be pigmented” (7–17). But what is the cause of this variability, the so-called “white Addison's disease” (16, 17)?

First, the pigmentation in pAI appears to be due to the binding of high levels of ACTH to the melanocortin 1 receptor (MC1R) (18). MC1R, a molecule with a high degree of sequence similarity to the MC2R, the ACTH receptor, binds to its regular ligand, α-melanocyte-stimulating hormone (α-MSH), and ACTH with an almost equal affinity (18). Levels of α-MSH in humans are typically low, but α-MSH stimulates melanogenesis in cultured human melanocytes and acts specifically to increase the synthesis of eumelanin (19). Both α-MSH and ACTH are splice products of the same precursor pituitary protein, proopiomelanocortin (POMC) (20), and appear to play a role in determining pigmentation in humans because homozygous or compound heterozygote mutations in the POMC gene are associated with hypopigmentation and red hair (21). In pAI (that is not properly replaced), POMC expression increases, and both α-MSH and ACTH increase; β-lipotropin, another POMC product, may also interact with MC1R and induce pigmentation in states of excess POMC synthesis (22).

The important role of MC1R in the determination of human skin and hair pigmentation is also beyond doubt: reduced function MC1R alleles leads to red hair, freckling, sun sensitivity, and an increased risk of cutaneous cancers, including melanoma (23–25). MC1R alleles accounted for varying pigmentation in Neanderthals, but the mutations were not the same as in modern humans, indicating that current MC1R-inactivating alleles have developed relatively recently and that there might be evolutionary pressure and/or advantage for their presence (26). Clearly, although MC1R is not the only determinant of skin color in humans (27) or other mammals (28), it is a major one.

Familial glucocorticoid deficiency (FGD) due to ACTH resistance consists of at least three distinct genetic syndromes that are all inherited as autosomal recessive traits: inactivating mutations of the ACTH receptor (the MC2R gene) and its accessory protein (MRAP); isolated ACTH resistance without MC2R, MRAP mutations, or any other defects in known genes; and Allgrove syndrome (AS) (29). In this issue of JCEM, Turan et al. (30) describe a patient with FGD without pigmentation: she was born with red hair which gradually darkened during early childhood; despite repeated episodes of hypoglycemia she was not diagnosed with pAI and high ACTH levels until she was 6 yr old. MC2R sequencing showed a homozygous T152K mutation that is known to affect trafficking of the receptor (31), like most ACTH receptor defects causing FGD. MC1R sequencing showed a homozygous R160W mutation that is among the most common genetic variants of the receptor in red-haired individuals (32). In a fascinating confirmation of the roles of ACTH (and possibly other POMC-derived peptides) and MC1R in determining not only skin but also hair pigmentation, the patient's hair “reverted to a reddish color” after proper replacement with hydrocortisone and a decrease of serum ACTH levels (30).

This is the first demonstration of MC1R genetic variants affecting the phenotype in pAI. It is possible that MC1R, the major freckle gene (25), modifies the phenotype in Carney complex (33) and other disorders associated with skin pigmentation and endocrine disease (i.e. McCune-Albright syndrome, neurofibromatosis syndromes, and others). Beyond the obvious diagnostic implications, MC1R variants also predispose to nevi and cancer (18, 23). For example, could the nevi observed in Turner syndrome and their response to GH (34) depend on MC1R function? And how does variant MC1R signaling affect the function of other MCRs that to some extent interact and coordinate their cAMP responses? Would the clinical presentation of an MRAP defect causing FGD or an AAAS mutation causing AS be affected equally by MC1R variants? In our experience (35), patients with AS do not frequently become heavily pigmented, despite very high ACTH levels, indicating that AAAS function may already affect MC1R signaling (beyond that of MC2R).

This last observation indicates that additional factors determine pigmentation in response to high ACTH levels. I recently asked Dr. White how many of his patients with pAI (36) presented with hyperpigmentation; he responded that “only two thirds (12 of 18) who presented ‘off the street’ were hyperpigmented” (P. C. White, personal communication). Given the much lower frequency of MC1R biallelic defects in the general, “mixed” population of the United States, “white Addison's disease” is neither as rare as previously thought (7–17) nor, most likely, exclusively due to genetic variants of the melanocortin system.

With genetic variants responsible for even subtler effects (than, for example, red hair), it is a new world out there for the practicing clinician, a world that does not contradict Occam's razor. But one has to think at all times in a different way, incorporating systems biology information in clinical practice. When I propose paraphrasing Hickam's dictum, “a patient can have as many genetic variants as he damn well pleases,” it does not mean that all these sequence defects cause various diseases; quite the opposite. In fact, the evidence is that there is some redundancy and a tremendously complex molecular balance in human biology (4, 5). But clinicians have to incorporate genetics in their daily practice, and educators have to introduce molecular pathways and their genetic variability in their teaching of classic physiology, pathophysiology, and clinical medicine.