‘An apple a day keeps the doctor away’: Revisiting phlorizin and the humble origins of SGLT2 inhibition

Abstract

The saying ‘An apple a day keeps the doctor away’ is often cited as a charming piece of folk wisdom rather than a scientifically grounded health recommendation. Yet, its enduring popularity highlights how empirical observations can sometimes foreshadow later scientific discoveries. When examined through the lens of medical history and pharmacology, this proverb offers an interesting opportunity to reflect on how empirical observations may sometimes precede mechanistic understanding. In particular, the discovery of phlorizin from the apple tree and its later role in the development of sodium–glucose cotransporter 2 (SGLT2) inhibitors provides a compelling historical perspective on the origins of modern glucose-lowering therapy.

In 1835, French chemists first isolated phlorizin, a dihydrochalcone glycoside, from the bark of the apple tree (Malus domestica).¹ At the time, the compound attracted limited clinical interest, largely due to its bitter taste and the absence of a clear therapeutic application. Like many plant-derived compounds described during the 19th century, its physiological relevance remained uncertain.

This changed in 1885, when Joseph von Mering and Oskar Minkowski observed that administration of phlorizin induced pronounced glucosuria in experimental animals, independent of pancreatic damage.² This finding was remarkable because it demonstrated that urinary glucose excretion could occur independently of diabetes itself, challenging prevailing assumptions about glucose homeostasis. Intriguingly, just 1 year later—around 1886—the proverb ‘An apple a day keeps the doctor away’ appeared in popular literature.³ Although purely coincidental, this temporal coincidence is historically striking and illustrates how nutritional beliefs and pharmacological discoveries may occasionally intersect.

For decades, phlorizin-induced glucosuria remained primarily a physiological observation rather than a therapeutic strategy. Its mechanism of action was gradually elucidated in the mid-20th century, when studies localized its effect to the renal proximal tubule, where glucose reabsorption occurs. Phlorizin was later shown to inhibit sodium-dependent glucose transport across the tubular epithelium.⁴

The molecular breakthrough came in the 1990s with the identification of SGLT2 as the primary transporter responsible for reabsorbing approximately 90% of filtered glucose in the kidney.⁵ Although phlorizin itself proved unsuitable for clinical use—owing to poor oral bioavailability, rapid hydrolysis, lack of selectivity and gastrointestinal side effects related to intestinal SGLT1 inhibition—it provided the essential pharmacological template for the development of selective SGLT2 inhibitors.

This ultimately led to the gliflozins, which are now established therapies not only for type 2 diabetes but also for heart failure and chronic kidney disease, with benefits extending well beyond glycaemic control.^{6, 7}

The development of SGLT2 inhibitors also reflects a broader evolution in diabetes management. Early therapies were based on correcting systemic metabolic abnormalities through insulin replacement or stimulation of secretion. As understanding of glucose physiology advanced, treatment strategies progressively shifted towards targeting specific molecular pathways involved in glucose homeostasis. Metformin reduced hepatic glucose production, thiazolidinediones improved insulin sensitivity through PPAR-γ activation, and DPP-4 inhibitors and GLP-1 receptor agonists exploited incretin biology to achieve increasingly mechanism-based glucose control.

SGLT2 inhibitors represent a distinct conceptual shift within this therapeutic evolution. Unlike previous glucose-lowering agents that primarily modulate insulin secretion or sensitivity, or incretin signalling, gliflozins target renal glucose handling and act independently of insulin pathways. Remarkably, their greatest clinical impact now extends beyond glycaemic control, with major benefits in heart failure, chronic kidney disease and cardiovascular prevention. In this sense, phlorizin did not simply inspire another antidiabetic drug class—it foreshadowed a therapeutic paradigm in which metabolic drugs became foundational cardiovascular therapies.⁵

Apples have long been recognized for their modest metabolic benefits, particularly attenuation of postprandial glucose excursions. This effect is most plausibly explained by their soluble fibre content, especially pectin, which delays gastric emptying and glucose absorption.⁸

Although apples also contain phlorizin and related dihydrochalcones, these compounds are concentrated mainly in the bark, peel, leaves and unripe fruit rather than in the edible pulp typically consumed in significant amounts.^{9, 10} Moreover, phlorizin has poor oral bioavailability and is rapidly hydrolysed in the intestine, making clinically meaningful systemic SGLT2 inhibition from normal dietary apple consumption highly unlikely.

Any potential glucose-lowering effect of dietary phlorizin would therefore be expected to be modest and more plausibly related to transient intestinal effects rather than sustained renal SGLT2 inhibition. Consequently, it would be inappropriate to attribute the historical health reputation of apples primarily to phlorizin-mediated glycaemic control. The benefits associated with apple consumption in the 19th century were likely more immediate and observable—such as digestive regularity, relief of constipation and general nutritional value—rather than measurable effects on glucose metabolism.

Nevertheless, the story of phlorizin remains a fascinating example of how observations surrounding natural products can ultimately contribute to major therapeutic advances.

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22 (Alexander et al., 2021).¹¹

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.