The Plasma Concentration of MUC5B Is Associated with Clinical Outcomes in Paraquat-poisoned Patients

To the Editor:

Paraquat (N,N′-dimethyl-4,4′-bipyridinium dichloride), a nonselective contact herbicide, is a common cause of poisoning in developing countries. Until the recent ban, paraquat was the most common cause of fatal self-poisoning in Sri Lanka. Ingestion leads to the development of multiorgan failure, severe pulmonary fibrosis, and death (1, 2). Paraquat-induced pulmonary fibrosis partly results from the direct damage caused by oxygen free radicals as well as indirect injury caused by inflammatory cells and fibroblasts. Mortality rates remain greater than 50% in self-poisoned patients because of the lack of an efficient antidote and limited response to antiinflammatory or antioxidant treatment (1, 3).

The MUC5B (mucin 5B, oligomeric mucus/gel-forming) gene encodes respiratory tract mucin glycoproteins that play a role in mucociliary clearance and host defense (4). We have previously shown that a common, gain-of-function MUC5B promoter polymorphism (rs35705950) is the strongest risk factor for the development of idiopathic pulmonary fibrosis (IPF) (5, 6). Moreover, among patients with IPF, the gain-of-function MUC5B promoter variant was significantly associated with improved survival (7). At this time, the best predictor of survival after self-poisoning by paraquat is time from ingestion and plasma paraquat concentration, termed the severity index of paraquat poisoning (SIPP score) (8). The objective of the present study is to determine whether the concentration of plasma MUC5B is associated with mortality in patients poisoned by paraquat.

We obtained plasma samples from a previously described cohort of self-poisoned patients with paraquat toxicity in Sri Lanka (3, 8, 9). Briefly, patients ingesting paraquat were enrolled at 10 different Sri Lankan hospitals between April 2005 and January 2008. Blood was obtained from patients on admission or during the patient’s first hospital day, and plasma paraquat levels were measured.

MUC5B from human saliva was used to generate a standard curve. Dot blot analysis on the plasma sample was performed using S&S Minifold I dot blot filtration manifold (Schleicher and Schuell). Samples were diluted 1:100 and run in duplicate along with the standard curve, also run in duplicate. The protein samples were allowed to bind to polyvinylidene difluoride membrane, the membrane was blocked at room temperature in 5% milk in phosphate-buffered saline with Tween for 1 hour then incubated overnight at 4° with the MUC5B primary antibody (Santa Cruz Biotechnology) diluted 1:5000 with 2.5% milk. Secondary goat anti-rabbit IgG biotin conjugate in 2.5% milk (1:2000 dilution) was applied for 1 hour at room temperature followed by incubation with strep-horseradish peroxidase (Life Technologies). Blots were developed using the horseradish fluoroilluminescence detection protocol, using SuperSignal Dura Extended Duration Substrate (Pierce).

We used a multivariate Cox proportional hazards model for survival analysis. We used the baseline SIPP score, which is the product of the time since ingestion and the serum paraquat concentration at admission (8), as the independent paraquat variable. The time until death or final follow-up was modeled as a function of the natural log of the baseline SIPP score, the MUC5B concentration, and the interaction of the log of the baseline SIPP and MUC5B. Sex and age were found to be nonsignificant and are therefore not included in the model. We calculated hazard ratios for the risk for death.

Demographics, baseline characteristics, MUC5B measurements, and SIPP scores of the 89 cases are summarized in Table 1. Figure 1 illustrates the survival probability based on MUC5B quantity and the log of baseline SIPP. The curve for patients with a log of baseline SIPP measurement below the median stabilizes around 63% for patients with a higher-than-median MUC5B plasma concentration and around 56% for patients with a lower-than-median MUC5B plasma concentration. Patients with a higher baseline SIPP have significantly worse outcomes, with survival probabilities leveling off around 5% for those with a lower-than-median MUC5B. All 26 patients who had a higher-than-median baseline SIPP measurement and a higher-than-median MUC5B died within the first 3 months.

Table 1.

Demographics and Case Characteristics of Patients Self-poisoned by Paraquat

Age, yr
<20	14 (15.7%)
20s	40 (44.9%)
30s	23 (25.8%)
40s	4 (4.5%)
50s	5 (5.6%)
60+	3 (3.4%)
Sex
Female	22 (24.7%)
Male	67 (75.3%)
In-hospital death
No	42 (47.2%)
Yes	47 (52.8%)
Death within 3 mo
No	28 (31.5%)
Yes	61 (68.5%)
Sampling time from ingestion, h
Mean (SD)	9.5 (7.35)
Median	7.1
Range	0.9 to 38.5
N	89
Baseline severity index of paraquat poisoning
Mean (SD)	64.4 (235.80)
Median	9.3
Range	0.1 to 1,936
N	89
Geometric mean (95% confidence interval)	7.49 (4.68 to 11.99)
MUC5B protein concentration in μg/ml
Mean (SD)	−0.0866 (0.0332)
Median	0.076
Range	−0.021 to −0.202
N	89

Definition of abbreviation: MUC5B = mucin 5B.

Figure 1.

Paraquat self-poisoning survival by MUC5B (mucin 5B) and log of baseline severity index of paraquat poisoning (SIPP). The median values of MUC5B (−0.0759) and log of baseline SIPP (2.22) are used to distinguish the four groups.

The Cox proportional hazards model demonstrates a hazard ratio of 0.78 (95% confidence interval [CI], 0.64–0.95) for MUC5B, suggesting a lower risk for death with higher MUC5B, and 0.93 (95% CI, 0.56–1.56) for the log-baseline SIPP score. The interaction of the plasma concentration MUC5B and log-baseline SIPP has a hazard ratio of 1.11 (95% CI, 1.04–1.19), suggesting that among patients with a low baseline SIPP, those with a higher quantity of MUC5B have higher chances of survival than those with a lower MUC5B. Among the relatively few patients with a high baseline SIPP, those with a higher concentration of MUC5B have lower chances of survival. Overall, high MUC5B concentrations improve prognosis except in those with very high SIPP scores. This suggests that higher paraquat concentrations may simply overwhelm the protective effects of MUC5B. In the future, we will seek to confirm these observations in animal models through overexpression and knock-out approaches.

Our findings demonstrate that MUC5B concentrations in the peripheral blood are associated with mortality in patients self-poisoned by paraquat. The amount of paraquat ingested and the time to presentation, as reflected by the SIPP score, remains the most predictive factor for subsequent mortality (8). However, the interaction of SIPP baseline score with MUC5B concentration suggests MUC5B may provide some protection from paraquat-induced mortality. These data paired with the prior knowledge that the MUC5B promoter variant rs35705950 is associated with improved survival in IPF (6) suggest that the MUC5B protein may be protective in oxidative injury leading to pulmonary fibrosis.

The persistently high mortality rate in paraquat self-poisoning (3, 10) highlights the need for improved treatment approaches. Pulmonary mucins may be a future therapeutic target for these patients. In addition, MUC5B may be a prognostic tool in poisoned patients with known timing and paraquat concentrations.

Lack of pulmonary imaging in this study limits our ability to associate the MUC5B concentrations with acute pulmonary toxicity, although mortality is arguably the most clinically important variable. We were unable to genotype these subjects because the purified plasma specimens contained very few peripheral blood mononuclear cells, which increased the difficulty of DNA extraction. Finally, lack of long-term follow up for pulmonary toxicity in the survivors minimizes our ability to comment on the development of paraquat-associated pulmonary fibrosis and the role of MUC5B in this pathology.