A POTENTIAL ROLE FOR IMMUNE ACTIVATION IN HEMODIALYSIS HYPOTENSION

Abstract

The necessary exposure of blood to biomembranes during hemodialysis has been viewed by many as an immunogenic challenge leading to an acute phase response. In this study we examined the relationship between hemodialysis-induced immune activation and intradialytic hypotension, using the acute phase reactant serum C-reactive protein (CRP) as a surrogate for immunogenic activation. The maximum percent change in mean arterial pressure (MAP) was found to correlate significantly with CRP (r = 0.67, p < 0.05) in nine consecutive patients with a history of symptomatic hypotension during hemodialysis. In contrast, no correlation was found between CRP and maximum percent change in MAP in eight consecutive hemodialysis patients without intradialytic hypotension. Since interleukin-6 (IL-6) is a major regulator of CRP, the relationship between these two proteins was examined. Plasma IL-6 levels were found to correlate both with CRP (r = 0.67, p < 0.05) and with mean maximum percent change in MAP (r = 0.70, p < 0.05) in hemodialysis patients with a prior history of hypotension. IL-6 levels did not correlate with CRP or blood pressure in the hemodynamically stable patients. The results suggest that immune activation working through IL-6, CRP and other cytokines may play a role in the pathogenesis of hemodialysis hypotension in some patients.

INTRODUCTION

Hemodialysis has been increasingly viewed as an immunogenic challenge that results in a recurrent acute phase response. This immunogenic challenge has occurred despite improvements in the “biocompatibility” of dialysis membranes. Footprints of immune activation are not infrequently observed in hemodialysis patients and include elevated levels of positively regulated acute phase reactants, such as serum amyloid A, C-reactive protein (CRP), fibrinogen, and Von Willebrands antigen [1-2], and/or reduced levels of negatively regulated reactants, such as albumin, and transferrin [[3]]. The inference from the results of these prior studies of immune activation is that the reaction ultimately plays a role in intradialytic hypotension, dialysis amyloidosis and malnutrition, to name a few common sequelae. The serial measurements of acute phase proteins in uremic patients reveals a significant increase in the levels of CRP after hemodialysis is begun [4-5], but a precise role of CRP in vivo has not been defined. However, in uremic patients on hemodialysis [[6]] and in the general population CRP [7-8] has been linked to fatal cardiovascular events. As is true of most acute phase proteins, CRP demonstrates tissue-specific expression in the liver and is regulated primarily by proinflammatory cytokines, such as interleukin-6 (IL-6), IL-1 and tumor necrosis factor (TNFα) at the level of gene transcription [9-12]. Since CRP has a half-life of approximately 18 h, we hypothesize that in some uremic patients, hemodialysis leads to the repetitive, cyclical production of proinflammatory cytokines, which maintains high plasma levels of CRP. Two observations support this hypothesis: (i) some patients on hemodialysis exhibit elevated serum levels of IL-6, IL-1β, and TNFα [[13]], [[15]]; and (ii) monocytes from some patients undergoing hemodialysis spontaneously release IL-1 in vitro and exhibit increased production of IL-1β, IL-6, and TNFα following stimulation with suboptimal doses of lipopolysaccharide (LPS) [16-20]. Taken together, the data are consistent with the interpretation that in uremic patients, CRP is a marker of recurrent, hemodialysis-mediated immune activation and, in those hemodialysis patients who display increased levels of CRP, hemodialysis-induced immune activation is accompanied by the release of proinflammatory cytokines which drive the acute phase response [[13]], [18-20].

As noted above, the same proinflammatory cytokines, which orchestrate CRP production, have been implicated in the pathogenesis of hemodialysis hypotension. Intradialytic hypotension is one of the most common debilitating complications of dialysis and is present in about 20 to 25% of hemodialysis encounters [[17]], [21-22]. The relationship between the proinflammatory cytokines and blood pressure is not fully understood but may depend on the ability of the cytokines to release nitric oxide (NO) in resistance vessels and thereby lead to hypotension. Hypotension during hemodialysis may be accompanied by decreased cardiac blood flow, which may lead to myocardial ischemia, decreased cerebral perfusion which may induce altered sensorium, fatigue, and headaches, and decreased oxygen delivery to skeletal muscles, which may result in claudication.

This study examines the potential relationship between the level of immune activation in hemodialysis patients and the occurrence of intradialytic hypotension. The results suggest that in some patients hemodialysis hypotension may be immunologically mediated.

METHODS

Patients

We hypothesized that in uremic patients, recurrent hemodialysis-induced immune activation results in, and is reflected by chronically elevated levels of CRP. Persistently elevated CRP, in turn, selects for patients that are prone to developing hypotension during hemodialysis. To test this thesis, we obtained CRP levels in nine consecutive patients receiving hemodialysis at the Medical College of Ohio outpatient dialysis facility, Toledo, Ohio, who had a history of recurrent, symptomatic intradialytic hypotension (defined as a reduction in mean arterial pressure of ≥20 mm Hg [[23]]) and eight consecutive hemodialysis patients with no prior history of hypotension during dialysis. Patients who were receiving immunosuppressive agents, had a recent infection, a febrile illness, or an inflammatory disease which might potentially elevate CRP were excluded from the study. The demographics of the patient population are summarized below (Table 1). All patients were dialyzed against a F80 polysulfone membrane (Fresenious Medical Center, North America, Lexington, MA). Dialyzer reuse was performed using DR54 Fresenious Bleach, then formaldehyde, and then heat for 24 h. Membranes were used up to 16 times/patient. All patients in the study had a urea clearance on dialysis (KT/V) of = 1.4.

Table 1. Demographics of the Study Population

No.	Status	Age^+	HD^*	Race	Etiology	Sex
1	Hypotensive	30	69	A.A.^§	SLE^ϕ	F
2	Hypotensive	27	108	A.A.	Glomerulonephritis	F
3	Hypotensive	84	17	Caucasian	Unknown	M
4	Hypotensive	70	38	Caucasian	Diabetes Mellitus	M
5	Hypotensive	54	17	Caucasian	Diabetes Mellitus	F
6	Hypotensive	48	40	A.A.	Diabetes Mellitus	M
7	Hypotensive	72	72	A.A.	Diabetes Mellitus	F
8	Hypotensive	70	33	A.A.	Diabetes Mellitus	F
9	Hypotensive	51	34	Caucasian	Diabetes Mellitus	M
10	Normotensive	47	14	A.A.	Glomerulosclerosis	M
11	Normotensive	47	15	Caucasian	Acute tubular necrosis	M
12	Normotensive	48	33	A.A.	Hypertension	M
13	Normotensive	69	34	A.A.	Diabetes Mellitus	M
14	Normotensive	19	5	A.A.	Glomerulonephritis	F
15	Normotensive	52	32	A.A.	Hypertension	M
16	Normotensive	68	24	Asian	Diabetes Mellitus	F
17	Normotensive	80	4	A.A.	Obstructive uropathy	M

⁺Age at the time of study

^*Months on dialysis at the time of study

^§A.A., African American

^ϕThe patient's SLE (systemic lupus erythematosus) was clinically and serologically inactive.

Measurement of CRP

CRP was measured in patients' sera by a sandwich-ELISA technique (detection limit 200 ρg/mL) using a modification of the method of others [[5]]. Briefly, microtiter wells (Nunc, Rosklide, Denmark) were coated with anti-human CRP antibody (Sigma, St.Louise, MO) at 1 μg/mL for 16 h at 4°C, and blocked with 0.2% casein (Sigma) in PBS for 1 h at room temperature. Wells were then washed with PBS containing 0.05% Tween 20 (Sigma). One hundred μL of sera, diluted in PBS-Tween, or CRP standards were added and incubated for 1 h. Following three washes, wells were incubated with 0.1 mL of biotin-labeled anti-CRP antibody (5 μg/mL) for 1 h, washed three times, and incubated with horseradish peroxidase-strepavidin (Amersham, Arlington Heights, IL) for 1 h. After incubation with 0.2 mM 2.2′-azinobis (3-ethylbenzthiazoline-sulfonic acid) (ABTS, Sigma) and 2.5 mM H₂O₂, optical density of wells was read at 405 μm using a microplate reader (Molecular Devices, Menro Park, CA). The upper limit of normal for CRP using this assay is 0.8 mg/dL.

Quantification of IL-6

IL-6 was quantified in patients' plasma by ELISA using a kit from Biosource International (Amarillo, CA). The detection limit in this assay is 0.15 to 10 ρg/mL.

Data Analysis

Paired t-tests were used to determine the differences between means. p < 0.05 was considered significant for all comparisons. Regression analysis was performed using the program Statistica (Apple Computers, Cupertino, CA).

RESULTS

In hemodialysis patients with a history of recurrent intradialytic hypotension, we found that the maximum percent change in mean arterial pressure (MAP) during dialysis correlated with CRP (Figure 1, upper panel, r = 0.67, p < 0.05). In contrast, no relationship was found between CRP and maximum percent change in MAP in eight consecutive hemodialysis patients with no history of intradialytic hypotension (Figure 1, lower panel).

Figure 1. CRP correlates with maximum percent change in MAP in hypotensive hemodialysis patients. CRP was measured prior to dialysis in nine consecutive hemodialysis patients with a history of intradialytic hypotensive (upper panel) and eight consecutive hemodialysis patients with no prior episodes of hypotension during dialysis (lower panel). Data show the relationship between CRP and maximum percent change in MAP recorded during the particular dialysis session in which CRP values were obtained.

Since the CRP varies little in individual dialysis patients over months, we examined the relationship between the maximum percent change in MAP and the CRP in the above nine, historically hypotensive patients and eight stable patients during five separate dialysis sessions; two prior to, and two following the session in which the data shown in Figure 1 were collected. A significant correlation was found over multiple dialysis sessions (n = 42) between the CRP and the maximum percent change in MAP (Figure 2, upper panel, r = 0.60, p < 0.001) and the mean maximum percent change in MAP (Figure 3, upper panel, r = 0.83, p < 0.01) in our nine hypotension-prone patients. In contrast, no relationship was evident between CRP and the maximum percent change in MAP (Figure 2, lower panel) or the mean maximum percent change in MAP (Figure 3, lower panel) during multiple dialysis sessions (n = 37) in our eight stable dialysis patients.

Figure 2. The relationship between CRP and maximum percent change in MAP during multiple hemodialysis sessions. CRP versus maximum percent change in MAP is shown in patients with (upper panel) or without (lower panel) a history of intradialytic hypotension during multiple sessions of hemodialysis; two prior to, and two following the session for which the CRP values shown in Figure 1 were obtained. Data were available from 42 sessions in hypotension-prone patients and from 37 sessions for stable patients.

Figure 3. The relationship between CRP and mean maximum percent change in MAP in hypotensive and stable hemodialysis patients. Shown is the mean maximum percent change in MAP versus the CRP recorded over 42 sessions in the nine hypotension-prone hemodialysis patients (upper panel) and recorded over 37 sessions in the eight stable hemodialysis patients (lower panel).

The association between CRP levels and maximum percent change in MAP in patients prone to dialysis hypotension suggested the possibility that, in such patients hemodialysis stimulated the release of proinflammatory cytokines which induced the synthesis of acute phase reactants. Since IL-6 is a major regulator of CRP, the potential relationship between these two proteins was investigated. Plasma IL-6 levels were found to correlate both with CRP (r = 0.67, p < 0.05) (Figure 4, upper panel) and with mean maximum percent change in MAP (r = 0.70, p < 0.05) (Figure 5, upper panel) in our nine patients with a history of intradialytic hypotension. No such correlations were found in hemodynamically stable patients (Figures 4 and 5, lower panels).

Figure 4. Plasma IL-6 levels correlate with CRP in patients with a history of hemodialysis hypotension. Pre-dialysis plasma IL-6 levels and serum CRP levels (drawn during the same dialysis session) were measured in patients with a history of intradialytic hypotension.

Figure 5. Plasma IL-6 levels correlate with mean maximum percent change in MAP in patients with a history of hemodialysis hypotension. Pre-dialysis plasma IL-6 levels are shown in relation to the maximum mean percent change in MAP for the nine hypotension–prone hemodialysis patients from 42 dialysis sessions.

DISCUSSION

The rationale for this study is based on the possibility that an unexplained elevation in CRP in patients on hemodialysis may be a marker of recurrent, dialysis-mediated immune activation and that this may contribute to subsequent hemodynamic instability on dialysis. We found that CRP levels correlated significantly with changes in MAP in hemodialysis patients with a history of intradialytic hypotension. This relationship was found to exist in patients over multiple dialysis sessions. In contrast, no association between CRP and MAP was evident in patients who were consistently normotensive during hemodialysis.

The elevated CRP levels in dialysis patients prone to hypotension suggested the possibility that hemodialysis induces the release of proinflammatory cytokines, such as IL-6, IL-1β and TNFα which promote the synthesis of acute phase reactants and cause local NO release. Consistent with this premise is the finding of a significant correlation between plasma IL-6 levels and CRP in patients with hemodialysis hypotension. It must be acknowledged that the number of subjects (and, therefore, data points) in this study is small, and, in the case of correlations dependent upon one or two points. However, the fact that there was some segregation in these data between hypotensive and non-hypotensive patients is intriguing, and suggests that a relationship between the cytokine and blood pressure is a tenable hypothesis.

Cytokines, such as IL-6, which regulate CRP activate inducible nitric oxide synthase (iNOS) and induce the production of NO [24-26]. A role for NO in hemodialysis hypotension was first suggested on theoretical grounds by Beasley and Brenner [[27]]. Noris et al. [[28]] subsequently showed that NO synthesis was increased in the platelets of uremic patients. More recently, Yokokawa et al. [[29]] reported that plasma levels of NO were higher in individuals who exhibited hypotension during dialysis than in those that remained normotensive during the procedure. The source of NO which is theorized to provoke hypotension in hemodialysis patients is postulated to derive from the activity of an inducible NOS, possibly present in vascular smooth muscle cells [[27]], platelets [[28]], endothelial cells [[30]], and/or leukocytes [[31]]. The work of Kang et al. [[32]] shows that NO is significantly cleared during dialysis and while inhibitors of inducible NOS are also eliminated intradialytically, preliminary data in our laboratory indicate that in most patients systemic, post-dialysis NO levels appear to be so low as to be biologically insignificant (Tomita et al. unpublished observation). These observations argue against a direct role for monocyte-derived NO in the pathogenesis of hemodialysis hypotension but do not exclude a potential contribution of local NO production by vascular smooth muscle cells to the pathogenesis of intradialytic hypotension, as postulated by Beasley and Brenner [[27]]. Since certain of the cytokines (e.g. IL-1β, TNFα), which upregulate gene transcription of both CRP and of iNOS are thought to be produced by intradialytic activation of mononuclear cells [4-6], in patients prone to experiencing dialysis hypotension, immune activation may result in the release of proinflammatory cytokines that promote hepatic synthesis of CRP and vascular smooth muscle production of NO. We are currently comparing the capacity of post-dialytic plasma from hypotensive versus non-hypotensive hemodialysis patients to stimulate NO release by smooth muscle cells.

Apart from NO, other vascular factors that provide potential partial explanations for hemodialysis hypotension include a reduction in the levels of the vasoconstrictor endothelin-1 [[33]], and an increase in the levels of the vasodilator adenosine [34-35]. An alternative, but not mutually exclusive hypothesis is that hypotension during hemodialysis is caused by NO derived from endothelial cells. Endothelial cells produce NOSIII, or eNOS and/or iNOS depending on their size [[36]]. eNOS is the major isoform found in microvascular endothelial cells (MVEC). Endothelial cells present in larger blood vessels, such as arterioles and arteries express both eNOS and iNOS [[36]]. The possibility that dialysis hypotension is mediated by local release of NO from MVEC has not been explored. Notwithstanding the fact that NO is constitutively produced by eNOS in MVEC [[37]], transcription of the gene, which encodes this enzyme is capable of being significantly upregulated by a variety of stimuli including fluid shear stresses [38-39], as might occur during hemodialysis. The shear stress response element present in MVEC has not been fully defined, and may consist of a number of cis-acting elements which bind distinct transcription factors, one of which may be nuclear factor kappa B [40-41]. The potential role of microvasculature in the pathogenesis of intradialytic hypotension is a fruitful area of future study.

In conclusion, the results of the present study show a correlation between hemodialysis hyopotension and CRP, which suggests that, in some hemodialysis patients, intradialytic hypotension may be immunologically mediated. The link between these immunologic pathways and hemodynamic stability is thus a plausible theory worthy of further investigation.