Fecal Transplantation in Hematopoietic Transplantation

In this issue of the Journal of Clinical Oncology, Rashidi et al [citation] report the results of a randomized trial of fecal microbiota transplantation (FMT) in recipients of induction chemotherapy for acute leukemia or allogeneic hematopoietic cell transplantation (allo-HCT). One hundred patients were randomized to receive FMT from healthy donors as oral capsules vs placebo upon recovery from neutrophil nadir. The primary endpoint was the all-cause infection rate over the following 4 months. The primary endpoint was not met, but the trial is still very instructive on several points.

Severe perturbations to microbiome composition have been described in both allo-HCT, and to a lesser but still significant extent in AML induction.^1–4 Low fecal microbiota diversity is a reproducibly observed predictor of mortality after allo-HCT^{3, 5–7}, primarily due to graft-vs-host disease (GVHD) and infections. Given the meticulous care with which indwelling catheters are now implanted and manipulated, bloodstream infections in allo-HCT patients are most often mucosal-barrier-injury associated. These infections of gut origin⁸ are very often preceded by a distortion of the usually diverse microbiome community such that ~30–100% of sequencing reads from fecal samples in the days prior to an infection are mapped to the single organism that caused the infection.^{9, 10} Under normal circumstances, such domination events are prevented by the normal microbiome community, which has evolved mechanisms to exert colonization resistance against such pathogens as Enterococci and Enterobacteriaceae (e.g. E. coli, Klebsiella).

The major mechanisms of colonization resistance by microbes are production of molecules that are toxic to neighboring organisms (most famously penicillin) and competition for key nutrients or other features of ecological niches. While restoring FMT is an investigational approach for systemic infection outcomes, restoring colonization resistance by these means is well-established in the case of C. difficile, for which fecal transplants (or products derived from fecal donors) have been recently authorized to prevent recurrence^{11, 12} With respect to other infections, the microbiome-derived metabolite desaminotyrosine can induce type-I interferon responses that protect experimental mice from influenza,¹³ and allo-HCT patients with upper-respiratory viral infections are more likely to progress to lower-respiratory tract infections if they harbor low diversity in their bacterial microbiomes.¹⁴ Invasive fungal infections have also been linked with intestinal fungal mycobiome domination events.^{15, 16} More broadly, an intact gut microbiome has been linked with more robust immune reconstitution after HCT in experimental animals, a prior randomized trial of FMT in allo-HCT recipients, and in a large observational cohort.^17–20 In principle, the risk of any opportunistic infection could be lessened by augmenting immune reconstitution, or perhaps even by modulating responses to vaccines.²¹

Besides microbe-microbe interactions within the gut lumen, restoring a healthy microbiome is expected to improve gut-barrier function^22–24, for example through provision of butyrate and indole metabolites that reduce permeability. Therefore, the risk of infection from any organism that gains entry by crossing the GI tract could be hypothesized to decrease upon restoring a normal microbiome by FMT; this might be the case in food-borne illnesses and in viral infections that may have a gastrointestinal reservoir.^{25, 26}

Moreover, the intestinal microbiome produces a large output of small-molecule metabolites, many of which are bioactive and achieve micro- or millimilar concentrations in the lumen and in the systemic circulation. The best studied among these is butyrate, which modulates the metabolism of colonocytes, functions as a histone deacetylase inhibitor, and induces regulatory T cells in mouse intestines.^{27, 28} Primary bile acids are excreted into the intestine and subsequently processed by microbes into a diverse pool of secondary bile acids that are then reabsorbed and modulate T cell differentiation and organismal metabolism.^29–32 A third example is the production of tryptophan-derived metabolites by microbiota, including inosines, kynurenanine, and serotonin, which have pleitotrophic effects on physiology.^{22, 33} Although many of these mechanisms have been well-studied only in animal models, their common theme is that a physiologic/homeostatic state is promoted by signals from a normal gut microbiome. The host, under normal circumstances, actively cultivates this homeostatic community of microbes by establishing a favorable gut-lumen niche that is temperature- and pH-regulated, anaerobic, and replete with nutrients from both alimentation and host-derived mucus, and into which antimicrobial peptides and IgA are secreted to further sculpt the composition of the community. Intensive cancer therapy, however, not only disrupts these host-derived inputs by causing mucositis but is also accompanied by poor nutrition and exposure to antibiotics and other drugs that directly affect bacteria as well.³⁴

Since for most disease states, the key relevant strains or their bioactive metabolites³⁵ that elicit the desired host response have yet to be defined, most FMT trials seek to restore this homeostatic feedback loop with a bulk transplant of feces from a healthy donor. A key assumption in these studies is that enough strains with relevant functions are likely to be in the stool of most healthy donors. A tacit assumption is that if the right strains can engraft, they might not only confer benefit direct but might jumpstart the feedback loop and nudge the host-microbiome interactions back toward a theoretically more homeostatic salutary state.

In order for a single (or perhaps a few) FMT treatments to plausibly affect clinical outcomes over the course of months, a transient presence of strains in the GI tract would need to durably modulate host biology, or the transplanted strains would need to durably engraft. Rashidi et al [citation] observed sustained engraftment of a considerable number of donor-origin taxa at 1 and even 9-months following the treatment. Notably, sustained engraftment was achieved by an encapsulated oral treatment (obviating enema or endoscopic instillation) and without the use of “conditioning” or “priming” regimen such as oral vancomycin, which has been employed in several recent and ongoing trials.^36–38

Although trial did not meet its prespecified primary endpoint, the study intervention did improve two key microbiota endpoints that could be considered surrogates or risk factors as they themselves have been previously well-correlated with mortality outcomes. First, fecal microbiota α-diversity was increased in FMT recipients. This next-generation-sequencing-assessed biomarker has predicted overall survival following allo-HCT in multiple observational cohorts.^{3, 6, 7} Second, the study treatment was able to mitigate expansion of Enterococci, which is not only a risk factor for Enterococcal bloodstream infections⁹ and mortality in leukemia,⁴ but exacerbates GVHD in mice and predicts for higher rates of GVHD and of mortality in multiple observational cohorts.³⁹

Of note, one design feature of the trial may have set the bar for success relatively high: The types of infections with the strongest evidence that they arise from microbial dysbiosis – C. difficle and mucosal-barrier-injury bloodstream infections – occur most frequently during neutropenic nadir which coincides with the period of greatest gut barrier damage. In this trial, the FMT was administered after neutrophil recovery/engraftment (also past the time of the initiation of alloreactivity which occurs in the first week post HCT). Thus the primary endpoint was more dependent types of infections that are less-clearly linked to the gut microbiome, such as viral (which comprised 33 of the 102 infection events in the whole study) and respiratory infections (also 33 of 102 events).

Another notable finding in the study was the overall safety of the FMT, even in these highly immunocompromised patients. Despite a prior report of a fatal case of transmission of multi-drug resistant pathogen via FMT to an HCT recipient⁴⁰, this study expands considerably the otherwise good safety track record of FMT in immunocompromised patients treated for hematologic malignancy with chemotherapy and allo-HCT.^{38, 41–47} Although there was a higher rate of GVHD in the recipients of FMT, there only 16 GVHD events in the whole study. Combined with an imbalance in the GVHD-prophylaxis regimens administered on the two arms, this make it difficult to interpret any effect on GVHD in this study, especially as FMT has demonstrated treatment responses in several studies of steroid-refractory GVHD.^{42, 44, 45} and studies of this are ongoing.³⁸

Future studies of FMT will need to study donor optimization, dose, route, timing, patient selection.^{36, 48} Antibiotic stewardship and evidence-based use of antibiotics and nutrition during treatment can help mitigate microbiome damage.⁴⁹ Finally, another approach that might overcome some of the barriers to FMT (such as batch-to-batch heterogeneity, healthy-donor recruitment challenges) is combinations of strains that are individually cultivated under clinical-grade production conditions.^{50, 51}

Figure: Fecal transplantation in the interaction between a patient with cancer and their microbiome.

The mammalian host cultivates a colonic microbiome through the provision of niche factors (e.g. temperature, pH, food, mucus) that favor the growth of certain microbial organisms. The microbiome, in turn, ferments foods the host cannot digest independently and also provides various inputs to the immune system and organismal metabolism, and also exerts colonization resistance against pathogens. This homeostatic feedback loop is disrupted by chemotherapy and hematopoietic transplantation. Fecal microbiota transplantation is hypothesized to restore colonization resistance and nudge this pathway back toward homeostasis.