Bugs, Drugs, and Shrugs

“For every action, there is an equal and opposite reaction” is Newton’s third law of thermodynamics.

Do the laws of physics have an extrapolation to human biology?

Experimental evidence “proves” that IL-17 plays a critical role in the pathogenesis of inflammatory bowel disease, but blocking IL-17 makes Crohn’s disease worse.[1, 2] Only a subset of patients with melanoma benefit from checkpoint inhibitors[3] and only a subset of patients with any rheumatic disease benefit from a specific therapy. A scientist at the NIH using a rodent model can’t replicate the experimental results of a scientist at Harvard. The intestinal microbiome might provide the explanation for each of these seemingly disparate conundrums.

These enigmas are not each directly addressed by Manasson and a large, talented group of international colleagues in their report in this issue of A&R on the effect of biologics, specifically TNF inhibitors or anti-IL-17, on the gut microbiome [4]. Nonetheless, their manuscript supports the hypothesis that unintentional effects on the microbiome could be the solution to many puzzling therapeutic observations. The “shrug” in our title reflects being flummoxed, i.e. how many of us respond to experimental or clinical data that seem to defy our theoretical predictions.

As discussed in their manuscript [4], Manasson and colleagues are not the first to examine how biologic therapies influence the microbiome [5, 6]. A dysbiosis or alteration in the gut microbiome is characteristic of the majority of immune-mediated diseases [7]. And in many cases, successful biologic therapy has been associated with a “normalization” of the gut microbiome [5]. To appreciate the implications of the Manasson report, we first provide some background about the microbiome and its relevance to rheumatic diseases.

We each possess two genomes that influence our health. The genome which is most obvious is the one that we inherit from our parents. The technology to alter this genome is just now emerging. The second genome is acquired. It is defined by the microbial world that cohabits our bodies. If the importance of a genome could be determined by the number of unique RNA transcripts it produces, the second genome would be around 150 times more important than the first.[8] The second genome also is not easy to alter, although it is certainly easier to manipulate (e.g. by diet) than our mammalian genome. The concept of a second genome has led to the recognition that we are holobionts, a product affected by both genome one and genome two. To understand the efficacy of a pharmaceutical, we need to know not only how the medication affects genome one; we also need to know how it affects genome two. And we need to recognize the reciprocal, how does genome two affect the medication (see Figure 1)? Acetaminophen [9] and methotrexate [10] are two widely used medications which are both metabolized by the microbiome. Even if a medication is not taken orally, the microbiome can still impact its efficacy. For example, the gut microbiome has a marked effect on the success of monoclonal antibodies in the form of checkpoint inhibitors in animal models of cancer [11, 12] as discussed below.

Figure 1.

The drawing indicates a positive feedback loop. The microbiome alters the effects of biologics, while biologics are altering the effects of the microbiome. The confused figure in the center of the drawing is puzzling to understand the implications of these interactions and wondering if many seemingly disparate observations could each be explained by changes in the microbiome.

The limited reproducibility of basic research has been widely hailed as a failing of modern science [13]. But what if the bacteria that dwell inside a mouse housed in St. Louis are vastly different from the bacteria living in a mouse from New Haven? Unlikely, you scoff. But it turns out that segmented filamentous bacteria (SFB) that live in the mouse gut have a profound effect on the immune system [14]. And if you purchase your C57Bl/6 mice from Taconic Farms, they are unlikely to harbor SFB, but the opposite is true if the mice are purchased from another major research animal vendor, Jackson Labs [15]. Differences in the microbiome that result from vendor, climate, diet, bacteria from the skin of a caretaker in a vivarium, or any one of limitless variables potentially could explain many examples of rodent research that suffer from poor reproducibility.

Perhaps the most direct extrapolation from the study by Manasson and collaborators relates to the explanation as to why blocking IL-17 does not ameliorate Crohn’s disease. Data from animal models shows a prominent role for IL-17 in several mouse studies of bowel inflammation [16]. The genetics of Crohn’s disease implicates the IL-23 receptor [16], and IL-23 is a major driver for IL-17 synthesis [17]. Immunohistology from human tissue supports the importance of IL-17 in the pathogenesis of Crohn’s disease [16]. Considering this strong rationale, many were shocked when inhibiting either IL-17A [1] or the receptor for IL-17 [2] actually exacerbated Crohn’s disease. Recent data point to a role for yeast or fungi, especially Malassezia, in the pathogenesis of Crohn’s disease [18]. IL-17 plays a major role in the control of fungal infections [19]. The data provided by the study in this issue indicate that blocking IL-17 increases the growth of Candida in the bowel, and thus could explain why Crohn’s disease is exacerbated by inhibition of IL-17 [20].

While it is not difficult to understand how bacteria could metabolize a medication taken orally, it is a greater challenge to theorize how the intestinal microbiome could alter the efficacy of a monoclonal antibody. However, bacteria in the gut seem to have a dramatic effect on the benefit of checkpoint inhibitors which are being used to bolster the immune response as a treatment for a variety of cancers [21]. The evidence for this effect is strong enough to support the rationale to study fecal microbiota transplants as a means to enhance the success of checkpoint inhibitor therapy [22].

As such, the report from Manasson and colleagues in this month’s A&R is a welcome addition to our understanding of the actions of two classes of biologics, those that inhibit TNF and those that neutralize IL-17A. Since both TNF and IL-17A are vital components of our immune system, neutralizing either would be expected to impact the microbiome. It behooves the rheumatologic community to define the impact on the microbiome from all medications including biologics. Is it all an epiphenomenon that we can largely ignore? Or is the alteration of the microbiome influential in determining who benefits from a medication and who does not? Who requires or tolerates an increased dosage? Who is likely to suffer an adverse effect?

Undoubtedly the greatest reward from biologics has derived from their impact on disease. But an unanticipated benefit has come from the lessons that have resulted from the adverse effects of this class of medications. For example, the association between TNF inhibitors and demyelinating disease has forced biomedical scientists to gain a better understanding of the role of TNF in the nervous system [23].

Isaac Newton’s laws have been applied to physics, but they also relate to biology. We are grateful to Manasson and colleagues for helping us better understand the multitude of unanticipated effects from biologic therapies.