BET Proteins Connect Diabetes and Cancer
Larry H. Bernstein, MD, FCAP, Curator
LPBI
New Proteins Discovered That Link Obesity-Driven Diabetes to Cancer
http://www.dddmag.com/news/2016/03/new-proteins-discovered-link-obesity-driven-diabetes-cancer
Killer T cells surround a cancer cell. Credit: NIH
For the first time, researchers have determined how bromodomain (BRD) proteins work in type 2 diabetes, which may lead to a better understanding of the link between adult-onset diabetes and certain cancers.
The findings, which appear in PLOS ONE, show that reducing levels in pancreatic beta cells of individual BRDs, called BET proteins, previously shown to play a role in cancer, may also help patients who are obese and diabetic.
The research was led by Gerald V. Denis, PhD, associate professor of pharmacology and medicine at Boston University School of Medicine, who was the first to show that BET protein functions are important in cancer development.
Adult-onset diabetes has been known for decades to increase the risk for specific cancers. The three main members of the BET protein family, BRD2, BRD3 and BRD4, are closely related to each other and often cooperate. However at times, they work independently and sometimes against each other.
According to the researchers new small molecule BET inhibitors have been developed that block all three BET proteins in cancer cells, but they interfere with too many functions.
“The BET proteins provide a new pathway to connect adult-onset diabetes with cancer, so properly targeting BET proteins may be helpful for both,” explained Denis, who is the corresponding author of the study.
He believes this discovery shows the need for deeper analysis of individual BET proteins in all human cell types, starting with boosting insulin and improving metabolism in the pancreas of adults who are obese.
“Without better targeted drugs, some ongoing cancer clinical trials for BET inhibitors are premature. These new results offer useful insight into drug treatments that have failed so far to appreciate the complexities in the BET family.”
Epigenetic modulation of type-1 diabetes via a dual effect on pancreatic macrophages and β cells
I-BET151 induces a regulatory phenotype in the pancreatic macrophage population
BET-protein inhibition promotes regeneration of NOD β cells
The studies presented here showed that treatment of NOD mice with the epigenetic modifier, I-BET151, for a mere 2 weeks prevented the development of NOD diabetes life-long. I-BET151 was able to inhibit impending insulitis as well as clear existing islet infiltration. The drug had a dual mechanism of action: it induced the pancreatic MF population to adopt an anti-inflammatory phenotype, primarily via the NF-κB pathway, and promoted β-cell proliferation (and perhaps differentiation). These findings raise a number of intriguing questions, three of which we address here.
First, why do the mechanisms uncovered in our study appear to be so different from those proposed in the only two previous reports on the effect of BET-protein inhibitors on autoimmune disease? Bandukwala et al. found that I-BET762 (a small-molecule inhibitor similar to I-BET151) altered the differentiation of Th subsets in vitro, perturbing the typical profiles of cytokine production, and reducing the neuropathology provoked by transfer of in-vitro-differentiated Th1, but not Th17, cells reactive to a peptide of myelin oligodendrocyte glycoprotein (Bandukwala et al., 2012). Unfortunately, with such transfer models, it is difficult to know how well the in vitro processes reflect in vivo events, and to distinguish subsidiary effects on cell survival and homing. Mele et al. reported that JQ1 primarily inhibited the differentiation of and cytokine production by Th17 cells, and strongly repressed collagen-induced arthritis and experimental allergic encephalomyelitis (Mele et al., 2013). However, with adjuvant-induced disease models such as these, it is difficult to discriminate influences of the drug on the unfolding of autoimmune pathology vs on whatever the adjuvant is doing. Thus, the very different dual mechanism we propose for I-BET151’s impact on spontaneously developing T1D in NOD mice may reflect several factors, including (but not limited to): pathogenetic differences in induced vs spontaneous autoimmune disease models; our broader analyses of immune target cell populations; and true mechanistic differences between T1D and the other diseases. As concerns the latter, it has been argued that T1D is primarily a Th1-driven disease, with little, or even a negative regulatory, influence by Th17 cells (discussed in [Kriegel et al., 2011]).
Second, how does I-BET151’s effect, focused on MFs and β cells, lead to life-long protection from T1D? MFs seem to play a schizophrenic role in the NOD disease. They were shown long ago to be an early participant in islet infiltration (Jansen et al., 1994), and to play a critical effector role in diabetes pathogenesis, attributed primarily to the production of inflammatory cytokines and other mediators, such as iNOS (Hutchings et al., 1990; Jun et al., 1999a, 1999b; Calderon et al., 2006). More recently, there has been a growing appreciation of their regulatory role in keeping diabetes in check. For example, the frequency of a small subset of pancreatic MFs expressing the complement receptor for immunoglobulin (a.k.a. CRIg) at 6–10 weeks of age determined whether or not NOD diabetes would develop months later (Fu et al., 2012b), and transfer of in-vitro-differentiated M2, but not M1, MFs protected NOD mice from disease development (Parsa et al., 2012).
One normally thinks of immunological tolerance as being the purview of T and B cells, but MFs seem to be playing the driving role in I-BET151’s long-term immunologic impact on T1D. Chronic inflammation (as is the insulitis associated with T1D) typically entails three classes of participant: myeloid cells, in particular, tissue-resident MFs; lymphoid cells, including effector and regulatory T and B cells; and tissue-target cells, that is, islet β cells in the T1D context. The ‘flavor’ and severity of inflammation is determined by three-way interactions amongst these cellular players. One implication of this cross-talk is that a perturbation that targets primarily one of the three compartments has the potential to rebalance the dynamic process of inflammation, resetting homeostasis to a new level either beneficial or detrimental to the individual. BET-protein inhibition skewed the phenotype of pancreatic MFs towards an anti-inflammatory phenotype, whether this be at the population level through differential influx, efflux or death, or at the level of individual cells owing to changes in transcriptional programs. The ‘re-educated’ macrophages appeared to be more potent at inhibiting T cell proliferation. In addition, it is possible that MFs play some role in the I-BET151 influences on β-cell regeneration. The findings on Rag1-deficient mice ruled out the need for adaptive immune cells in the islet infiltrate for I-BET151’s induction of β-cell proliferation, but MFs are not thought to be compromised in this strain. Relatedly, the lack of a consistent I-BET151 effect on cultured mouse and human islets might result from a dearth of MFs under our isolation and incubation conditions (e.g., [Li et al., 2009]). Several recent publications have highlighted a role for MFs, particularly M2 cells, in promoting regeneration of β cells in diverse experimental settings (Brissova et al., 2014; Xiao et al., 2014), a function foretold by the reduced β-cell mass in MF-deficient Csf1op/op mice reported a decade ago (Banaei-Bouchareb et al., 2004).
Whether reflecting a cell-intrinsic or -extrinsic impact of the drug, several pro-regenerative pathways appear to be enhanced in β-cells from I-BET151-treated mice. Increased β-cell proliferation could result from up-regulation of the genes encoding Neurod1 (Kojima et al., 2003), GLP-1R (De Leon et al., 2003), or various of the Reg family members (Unno et al., 2002; Liu et al., 2008), the latter perhaps a consequence of higher IL-22R expression (Hill et al., 2013) (see Figure 6B and Supplementary file 4). Protection of β-cells from apoptosis is likely to be an important outcome of inhibiting the NF-κB pathway (Takahashi et al., 2010), but could also issue from enhanced expression of other known pro-survival factors, such as Cntfr (Rezende et al., 2007) and Tox3 (Dittmer et al., 2011) (see Figures 4 and 6B). Lastly, β-cell differentiation and function should be fostered by up-regulation of genes encoding transcription factors such as Neurod1, Pdx1, Pax6, Nkx6-1 and Nkx2-2. The significant delay in re-onset of diabetes in I-BET151-treated diabetic mice suggests functionally relevant improvement in β-cell function. In brief, the striking effect of I-BET151 on T1D development in NOD mice seems to reflect the fortunate concurrence of a complex, though inter-related, set of diabetes-protective processes.
Lastly, why does a drug that inhibits BET proteins, which include general transcription factors such as Brd4, have such circumscribed effects? A 2-week I-BET151 treatment might be expected to provoke numerous side-effects, but this regimen seemed in general to be well tolerated in our studies. This conundrum has been raised in several contexts of BET-inhibitor treatment, and was recently discussed at length (Shi and Vakoc, 2014). The explanation probably relates to two features of BET-protein, in particular Brd4, biology. First: Brd4 is an important element of so-called ‘super-enhancers’, defined as unusually long transcriptional enhancers that host an exceptionally high density of TFs—both cell-type-specific and general factors, including RNA polymerase-II, Mediator, p300 and Brd4 (Hnisz et al., 2013). They are thought to serve as chromatin depots, collecting TFs and coordinating their delivery to transcriptional start-sites via intra-chromosome looping or inter-chromosome interactions. Super-enhancers are preferentially associated with loci that define and control the biology of particular cell-types, notably developmentally regulated and inducible genes; intriguingly, disease-associated, including T1D-associated, nucleotide polymorphisms are especially enriched in the super-enhancers of disease-relevant cell-types (Hnisz et al., 2013;Parker et al., 2013). Genes associated with super-enhancers show unusually high sensitivity to BET-protein inhibitors (Chapuy et al., 2013; Loven et al., 2013;Whyte et al., 2013). Second: although the bromodomain of Brd4 binds to acetyl-lysine residues on histone-4, and I-BET151 was modeled to inhibit this interaction, it is now known to bind to a few non-histone chromosomal proteins as well, notably NF-κB, a liaison also blocked by BET-protein inhibitors (Huang et al., 2009; Zhang et al., 2012; Zou et al., 2014). Abrogating specific interactions such as these, differing according to the cellular context, might be the dominant impact of BET inhibitors, a scenario that would be consistent with the similar effects we observed with I-BET151 and BAY 11–7082 treatment. Either or both of these explanations could account for the circumscribed effect of I-BET151 on NOD diabetes. Additionally, specificity might be imparted by different BET-family members or isoforms—notably both Brd2 and Brd4 are players in MF inflammatory responses (Belkina et al., 2013). According to either of these explanations, higher doses might unleash a broader array of effects.
Islet inflammation: A unifying target for diabetes treatment?