Page 18 - Read Online

Page 12 of 18 Mueller et al. Microbiome Res Rep 2024;3:33 https://dx.doi.org/10.20517/mrr.2024.09

We next asked if the relative abundance of A. muciniphila subspecies (AmIa and AmIb), influenced their
correlation with health. In the POMMS dataset, there was no significant difference in the relative abundance
of A. muciniphila AmIa (U = 7,574, P = 0.5084) or AmIb (U = 7,115, P = 0.2384) between case (obese) and
control samples [Figure 3D]. For the combined IBD dataset [Figure 3E], we found that there is a significant
difference in both AmIa [H(3) = 11.75, P = 0.0028] and AmIb [H(3) = 11.08, P = 0.0039] between healthy
and IBD groups. Post hoc pairwise comparisons indicated that AmIa is significantly decreased in both the
CD and UC groups (control mean rank = 192.8, CD mean rank = 177.9, P = 0.0178, UC mean rank = 173.8,
P = 0.0015), while AmIb is only significantly decreased in the UC group relative to healthy controls (control
mean rank = 192.4, UC mean rank = 163.3, P = 0.0121).

Akkermansia is associated with improved outcomes following PD-1 blockade treatment of epithelial
[6]
tumors and the relative abundance of Akkermansia is predictive of clinical responsiveness to similar
treatment of non-small-cell lung cancer (NSCLC) . In particular, the presence of Akkermansia in patient
[41]
stools was associated with improved responses to immunotherapy and overall survival, compared to
patients with no Akkermansia. We reanalyzed stool metagenomes from one of these studies and assigned a
[41]
dominant Akkermansia species and subspecies to each sample . Based on a Cox regression model, adjusted
for sex and age of patients [Figures 4A and B], we found a significant survival difference between patients
colonized predominantly by A. muciniphila (AmI) and those without Akkermansia (HR = 0.580, 95%CI:
0.408-0.820, P = 0.002). There was no significant survival difference between patients colonized
predominantly by A. massiliensis or A. biwaensis.

We also found a significant difference between A. muciniphila AmIa and AmIb in the survival of patients
with NSCLC undergoing immune checkpoint inhibitor treatment. After restricting our analysis to those
samples with detectable A. muciniphila, we generated a Cox regression model [Figures 4C and D] and found
that patients colonized predominantly by AmIa had a significant increase in survival compared to patients
with no detectable A. muciniphila (HR = 0.492, 95%CI: 0.319-0.744, P = 0.001). We did not detect this
enhanced survival in patients predominantly colonized with A. muciniphila AmIb (HR = 0.689, 95%CI:
0.434-1.062, P = 0.101).

Overall, our findings indicate that in circumstances where A. muciniphila is associated with an impact on
host physiology or immunity, which subspecies predominates can make a difference in its association with
disease outcomes. In IBD, AmIa abundance is linked to protection from both CD and UC, while AmIb was
only associated with protection from UC. Likewise, the increased survival noted in NSCLC patients is only
significant for those patients who were colonized by predominantly AmIa.

DISCUSSION
We performed a detailed analysis of genomic variance among human Akkermansia isolates and reinforced
the existence of at least six phylogroups (AmI-AmVI) [19-22] . We also provide further support for the
assignments of new Akkermansia species, A. massiliensis and A. biwaensis, and for at least two
A. muciniphila sub-species groups. These conclusions are based on meeting the threshold of separate
species as assessed by a 95% genome ANI threshold, which is further supported when comparisons are
limited to proteins encoded by Akkermansia core genes. Species assignments have traditionally used a 97%
threshold of 16S rRNA gene identity, and at a less restrictive threshold of 98.65%, only one new species
would be formed, containing the AmIV and AmVI phylogroups. However, discrepancies between ANI and
16S rRNA gene identities when making species classifications are not uncommon. Genome-based
classifications of Burkholderia and Anaplasma isolates have identified strains that rise to the level of new
species by ANI, but not 16S rRNA genes, suggesting that the utility of 16S rRNA identity as a sole