Patterns of gene expression in CD8+ T cells could predict prognosis and may help inform treatment decisions for patients with Crohn's disease and ulcerative colitis, according to the results of an early-stage study by researchers at the University of Cambridge.
The team, led by Kenneth Smith, a Cambridge immunologist, used gene expression profiling to identify a CD8 cell transcriptional signature that segmented patients with Crohn's or UC into two otherwise indistinguishable subgroups: one that experienced a much worse disease course than the other.
The ability to divide patients into two prognostic subsets could allow for more rapid treatment for those who need it, and the avoidance of serious side effects in those who don't, the group reported in the Journal of Clinical Investigation this month.
The researchers identified the signature using a population of 67 patients and are now planning a larger prospective clinical trial that will compare tailored treatment based on CD8 subtypes with the standard step-up approach currently used for most CD and UC sufferers. At the same time, they are working on adapting their method to profile unseparated blood samples rather than separated immune cell types, as well as investigating the signature's potential in other immune diseases and viral infections. They plan to work to develop a commercial prognostic assay.
Smith told PGx Reporter this week that his team had previously identified CD8 transcriptional signatures in two unrelated autoimmune diseases, lupus and antineutrophil cytoplasmic antibodies-associated vasculitis. "We wanted to see if the principle we'd established in autoimmune disease held true in diseases thought of as more auto-inflammatory," and it did, he said.
"We realized if we ever wanted to do a clinical trial we needed a disease that was more common and also where early therapy is known to be advantageous so you could use a biomarker to base an accepted therapy. All of that is true of IBS, particularly with Crohn's," he said.
The group has not limited its IBS study to CD8 cells, but thus far has found that only CD8 expression signatures are useful for prognosis, while other T cell and other immune cell types may offer adjunct information. Initially, the researchers expected CD4 cells, which evidence has suggested are implicated in the pathogenesis of both CD and UC, to offer some prognostic potential, but so far they have not found "a clear signature correlated with disease outcome," Smith said.
In their study, researchers tested a cohort of 35 Crohn's disease patients and 32 patients with active ulcerative colitis. About 60 percent were newly diagnosed, while the rest had had a period of disease regression after an initial diagnosis. The patients were then managed conventionally using a standard step-up strategy.
From 110 mL blood samples, the team isolated the various cell types, removing CD4+ and CD8+ cells for gene expression analysis using Affymetrix Human Gene ST microarrays.
Smith said that earlier studies attempting to discern prognostic signatures in un-separated cells have been disappointing because of the noise introduced as white cell subsets differ in proportion over time. "By separating the white cell subtypes you get a much better signal-to-noise ratio and see patterns emerging," he said.
Using unsupervised bioinformatic analysis, the researchers found that 3,403 CD8+ genes in the CD cohort and 4,186 genes in the UC cohort were significantly differentially expressed in two patient subgroups. Comparison of the signatures revealed high overlap between the two diseases — in fact, both signatures could be used interchangeably in each disease to discern between subtypes, according to the group's report.
The researchers then zeroed in on a smaller set of genes that could predict placement in one of the two subtypes, and termed the two subgroups IBD1, characterized by higher gene expression, and IBD2, characterized by lower expression.
Next they evaluated whether placement in the two groups correlated with clinical progression and disease severity in UC and CD. In the CD cohort, IBD1 patients experienced a worse disease course with greater relapsing and chronically active disease, and required treatment escalation. The UC cohort mirrored this. The researchers calculated the prognostic utility of the gene signature, reporting 89 percent specificity in CD and 84 percent in UC. Sensitivity was 59 percent in CD and 77 percent in UC.
"Prestratifying patients according to their IBD1/2 subgroup membership would have effectively distinguished those patients who were destined to experience more aggressive disease from those who went on to experience indolent disease," the authors wrote.
Smith said the team is still following up with patients and collecting detailed data on their disease course. The potential to be able to make better informed treatment decisions at the onset of either Crohn's or UC, he said, is exciting.
"At the moment because we can't differentiate [the severity of disease] when people show up for the first time, in order to do no harm the general approach is to minimize initial therapy and only escalate to immunosuppression and surgery in the patients that really need it," he said.
"This is particularly good for the group that doesn't need aggressive treatment. They are treated appropriately lightly. But the people who do need it are delayed and may end up with tissue damage or surgery they wouldn't have otherwise needed."
He added that there have been several studies "demonstrating that using anti-TNF-a therapy early actually increases its efficacy," Smith said. "The aim of this test would mean we could still treat the low group with low treatment but it will allow earlier treatment for the [poorer prognosis] group."
In a commentary published in the same issue of JCI, three independent researchers weighed in on some of the more far-reaching implications of Smith and his colleagues' findings.
The group's gene expression signatures, they wrote, "may help refocus research efforts to understand the pathogenesis of IBD and the mechanisms of treatment resistance."
Smith "liked the commentary," which he said put into words "many of the more speculative possibilities his group has avoided mentioning in their study."
Most of the pathways associated with the genes the group identified in the signatures are not really surprising, and several, like IL-7, already have drugs being developed to target them, Smith said. "Our results may well hint that many agents already being made might be useful in these diseases and may also tell you which subset of patients they will be useful in."
"The important caveat is that though we've seen this in CD8 T cells, it may not be causative," he said. "That doesn't matter in terms of therapeutic decision-making, but it would if you want to hit this as your therapeutic target."
Meanwhile, the researchers are planning to look further into why they weren't able to associate CD4+ gene expression with prognosis, when that cell type is thought to be more closely related with IBD.
"We think we do [have a sense of why CD4 didn't work] and are planning to publish that in the future," he said. "We don’t think [the fact that they were not prognostic] means CD4s aren’t playing a role. We just think it means they're more complicated."
Using what they know about the two CD8 subtypes, the researchers believe that further analyses will reveal more about the role of CD4 gene signatures. Smith said he didn't expect this to affect the prognostic power of the CD8 signature, which he said was "pretty much as good as it gets," but a better understanding of the distribution of CD4 gene expression could help the researchers better understand the disease mechanisms.
The group is also working now to move forward on a planned prospective trial looking at the CD8 signature. "We will use the test to determine treatment and randomize people; some will get the treatment normally given and some we'll use the test to determine, and then see who does better in terms of efficacy and toxicity," Smith said.
Separately, the researchers are working to develop a commercial test based on their research. Having secured relevant patents, Smith said, they are now examining other diseases to extend the breadth of the potential prognostic assay.
"Given that it affects both autoimmune and autoinflammatory disease, we should be thinking about all the other autoimmune diseases, but potentially also whether it predicts infectious disease outcomes," he said.
"We are also trying to make the test simpler, and to develop a test that works on whole blood" rather than requiring separated CD8 cells, he said. "For a lot of this — everything except the clinical trial — we expect to have made substantial progress within a year."
Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at mashford [at] genomeweb [.] com.