Markers or Makers: Inflammatory Cytokines in Treatment Resistant Hypertension

Between 7 to 20% of the hypertensive population exhibits resistance to treatment, defined as blood pressure that is poorly controlled with 3 drugs or requiring 4 drugs for control. The diagnosis of resistant hypertension requires exclusion of poor compliance, secondary causes and white coat effect. A possible cause of resistant hypertension that has been previously investigated in small studies is inflammation. In this issue of Hypertension, Chen et al report a previously undefined relationship between some biomarkers of inflammation and apparent treatment resistant hypertension (ATRH) in patients with chronic kidney disease.¹ The investigators analyzed 1,359 subjects with ATRH from the Chronic Renal Insufficiency Cohort and 2,008 subjects without ATRH. They found that resistant hypertension was associated with a rather marked increase in the inflammatory markers tumor necrosis factor-α (TNF-α) and Interleukin-6 (IL-6) (odds ratios 1.49 and 1.29 respectively) and a reduction in levels of Transforming growth factor beta-1 (TGF-β). In contrast, blood levels of the high sensitivity C reactive protein (hsCRP), fibrinogen, IL-1β and the IL1 receptor antagonist (IL-RA) were not associated with resistant hypertension. The authors created an inflammatory score, based on the levels of each of these biomarkers and found that higher scores were associated with resistant hypertension. Moreover, participants were followed for an average of 6.7 years and hazard ratios associated with ATRH were 1.49 for cardiovascular diseases (CVD) and 1.27 for death after adjustment for confounding variables.

Why would the Chen et al choose to study the role of inflammatory markers in ATRH? Studies during the past decade, largely in experimental animals, have established a role of the immune system in the development of hypertension and its attendant end-organ damage. It is now clear that cells of both the innate and adaptive immune system accumulate in kidneys and vessels of hypertensive animals and humans. These cells release inflammatory cytokines, including interleukin 17A, TNF-α, interferon gamma (IFN-γ), IL-6 and others that promote vascular and renal dysfunction and ultimately tissue damage.² Of relevance to the Chen et al study, Filho et al. have shown that neutralizing circulating TNF-α reduces blood pressure in Spontaneously Hypertensive Rats.³ Likewise, Guzik et al showed TNFα blockade lowers blood pressure and reduces vascular superoxide production in ang II-infused mice.⁴ Mice lacking IL-6 exhibit reduced hypertension and less reduction of renal blood flow during ang II infusion.⁵ Thus, the finding of elevated IL-6 and TNFα in patients with chronic kidney disease and ATRH are in keeping with prior studies in experimental animals.

Prior studies have examined the association of hsCRP, IL-6 and TNF-α with hypertension.^{6, 7} Like Chen et al, Bautista et al found a significant association between blood pressure and plasma levels of TNFα and IL-6, but no relationship to hsCRP in 196 otherwise healthy subjects⁶. hsCRP has been also been shown to be predictive of the development of hypertension⁸. In smaller cohorts of hypertensive subjects, Yu et al showed higher levels of TNF-α and IL-6 in hypertensive patients,⁷ while studies from our group showed higher TNF-α and hsCRP levels, but no differences in IL-6 levels in hypertensive and resistant hypertensive subjects compared to normotensive.⁹ The Chen et al study adds to these prior reports in several ways. First, it was is a study not simply of hypertension, but of resistant hypertension, a special subset that is of particular clinical interest. It is very rare to have such a large population of patients with resistant hypertension. Second, it involved patients with chronic kidney disease, a condition where inflammation is common. Among the subjects studied by Chen et al, those with resistant hypertension had higher urinary protein and significantly lower GFR as compared to the patients without resistant hypertension, possibly reflecting more severe kidney disease. Multivariate analysis showed a significant relationship between the inflammatory cytokines and resistant hypertension even after correcting for renal function. While drugs that modify immune responses would unlikely ever be used in humans with easily controlled hypertension, agents like anti-IL6 or anti-TNFα might be employed in a high-risk population such those with resistant hypertension.

At least 4 of biomarkers studied by Chen et al, IL-6, IL-1β, CRP and TNFα are intimately related (Figure 1). IL-1β can signal IL-6, which in turn signals production of CRP. Moreover, TGFβ and IL-6 act in concert to stimulate T cells to produce IL-17. TNFα can also stimulate production of CRP. It is therefore somewhat surprising that these three biomarkers did not track in parallel in the study of Chen et al., however there are other examples where these are not equally elevated or seem to provide different prognostic information. There are several potential reasons for discordance between biomarkers like IL-1β, IL-6 and CRP. One is that cytokines like IL-6, TNFα and IL-1β are produced locally and have very powerful effects on adjacent cells. The amount that enters the circulation may or may not reflect these local concentrations and little is known about factors that influence clearance of these from the blood. Thus, IL-1β might be elevated within the interstitium of the vessel or the kidney, but not changed in the circulation. IL-6 can be stimulated by other stimuli, like the Jak/STAT pathway, which in turn promotes further IL-6 production. Another very important issue relates to the assays involved. Biomarkers like IL-6 and TNFα are usually measured by enzyme-linked immunosorbent assays or cytokine bead arrays, often in individual laboratories. Such assays are notoriously fickle, and are not standardized between one laboratory and the next. There is little information as to stability of these cytokines in stored samples or the coefficient of variation of measurements. Normal IL-6 values have been reported to be less than 0.5, 1.5 and < 4 pg/ml in various studies. In normal subjects, circulating levels of IL-1β are often undetectable.⁹ Thus, the inability to detect a change in one biomarker might reflect inaccuracy of the bioassay or instability of the biomarker in stored samples. Alternate approaches are being used to characterize immune cells in humans, including analyzing the cytokines they produce after being placed in culture, using high density phenotyping like mass cytometry (CyTOF) and single cell sequencing. Another issue is that these biomarkers, like the sedimentation rate or leukocyte count, are not specific for any one disease. IL-6 levels are increased in response to exercise and in conditions like cancer, mental illness and asthma. Thus an elevation of a biomarker like IL-6 in any one individual could be due to coexisting conditions and must be evaluated within its clinical context. Anti-hypertensive drugs, similar to those taken by subjects in the study by Chen et al, might also decrease certain circulating cytokines.

Factors in the hypertensive milieu stimulate immune and parenchymal cells, including endothelial cells and renal cells to produce inflammatory cytokines. Activated immune cells infiltrate target organs, including the kidney and vasculature, leading to renal and vascular dysfunction. These factors not only worsen blood pressure elevation but also promote end-organ damage. Chen et al devised an inflammatory score that was elevated in subjects with treatment resistant hypertension and was associated with increased cardiovascular events.

Beyond the issue mentioned above, the finding that some cytokines, like TNFα and IL-6, but not others, like IL-1β and TGFβ are elevated in hypertension might seem inconsistent, because most of these cytokines are considered “pro-inflammatory”. It is critical to understand that not all inflammation is the same, a fact recognized by rheumatologists for years. An example is interleukin-17 (IL-17), which has been implicated in the pathophysiology of several autoimmune diseases including psoriatic arthritis, rheumatoid arthritis, ankylosing spondylitis, asthma and multiple sclerosis. While blockade of IL-17 has proven extremely useful in several of these, it is ineffective in Crohn’s Disease.¹⁰ Another example is blockade of IL-6, which is highly effective in rheumatoid arthritis but less effective in ankylosing spondylitis.¹¹ Two recent large clinical trials have highlighted this concept for treatment of cardiovascular disease. The CANTOS trial showed that targeting IL-1β in high-risk patients with previous history of myocardial infarction and high levels of hsCRP reduced cardiovascular events.¹² In contrast, low-dose methotrexate failed to reduce cardiovascular events and had no effect on levels of IL-1β, IL-6 and CRP in the CIRT study.¹³ The populations studied in these two trials had different levels of hsCRP and therefore their disparate results might be due to varying levels of baseline inflammation, however another explanation for their discordant results could differences in the actions of these drugs on the underlying pathology of atherosclerotic lesions. Taken together, these studies emphasize the importance of characterizing the immune process in a disease like hypertension and thereby identifying appropriate therapy. Studies like that of Chen et al are informative in beginning to address this issue.