Serum biomarkers predictive of cure in Chagas disease patients after nifurtimox treatment

Chagas disease (CD) is an important public health concern in Latin America, with an estimated eight million persons chronically infected and a total cost due to loss of productivity of ~$1.2 billion US/year [1] and $24.73 billion in health-care costs [2]. There are at least 41,000 new cases of CD due to vectorial transmission each year [3]. Regardless of the route of transmission (eg: triatomine insects, transfusion, mother-to-child, transplantation, or contaminated foods), the majority of individuals infected by T. cruzi remain asymptomatic for life [4]. However, 20-30% will eventually suffer cardiovascular, gastrointestinal, and/or neurological complications [1]. Once exclusively an ‘exotic’ disease, CD has become a burden in many non-endemic countries through migration, blood transfusion, vertical transmission and organ transplantation. Approximately 300,000 chronically-infected individuals are currently thought to be living in the United States [5]. Many European countries, Australia, Japan and Canada are experiencing similar problems. For example, Spain currently has >67,000 CD infected immigrants [6].

Despite the migration of large numbers of chronically-infected CD patients in non-endemic countries, CD remains a neglected disease. Since the introduction of Nfx in 1965 and benznidazole in 1971, the development of new drugs for CD has been negligible [7, 8]. Parasitological cure rates for benznidazole or Nfx vary with the disease phase, dose, age, geographical origin and treatment duration [9–12]. There is no consensus regarding optimal treatment duration which varies from 30 days to 90 days depending on the tolerance to the drugs. There is also no reliable test of cure that can be used to monitor the effectiveness of treatment in chronic CD patients [13].

Acute CD phase, which lasts for ~ ten weeks, is relatively easy to diagnose by microscopy, polymerase chain reaction (PCR), hemoculture or/and xenodiagnoses [14]. However, the vast majority of CD cases are not recognized and the disease enters the chronic phase when these tests targeting parasite material in the blood become unreliable due to the low and variable levels of parasitemia. Therefore, negative results using these tests cannot be used to either exclude infection or confirm parasite clearance. During this chronic stage, the diagnostic methods of choice are serologic including indirect immunofluorescence (IFA), indirect hemaglutination (IHA) or enzyme-linked immunosorbent assays (ELISA) [15–18]. Unfortunately, these serologic assays often yield inconclusive results [19] and the Pan American Health Organization (PAHO) now recommends the use of two different tests in parallel [20]. Although these tests can remain positive for decades after treatment, many investigators still rely on them to assess cure [18, 21–23]. In the last few years, several studies focusing on biomarkers of cure in chronic human disease have been proposed based on hemostatic parameters [24], natriuretic peptides [25], recombinant antigens [26], and lytic antibodies [27]. None of these approaches is particularly sensitive and new biomarkers to monitor responses to CD treatment are needed.

In recent years, mass spectrometric approaches have seen increased use for a wide range of infectious diseases [28–33]. We have used mass spectrometry to identify biomarkers for Fasciola hepatica in sheep [34], Taenia solium in pigs [35], and T. cruzi in chronically infected patients [36]. The purpose of the current study was to test our candidate T. cruzi biomarkers in a cohort of adult patients treated with Nfx and followed up after 3 years [21]. In this work, we not only found the same biomarker pattern in the subjects before treatment but confirmed their disappearance in almost half of the subjects after treatment. Based on these findings, we propose an algorithm to identify patients with chronic CD who have been cured by treatment.

The identification of T. cruzi by xenodiagnosis, hemoculture, microscopy or DNA based assays provides a definitive diagnosis for Chagas disease and can be routinely used to accurately diagnose the disease in the acute phase. However, during the indeterminate or chronic phase these tools are unreliable and a negative result may not mean lack of infection or parasitic cure of an infected patient, with serological methods being more informative.

Once a positive diagnosis has been confirmed, there are only two drugs available for treatment (Nfx and benznidazole) [38]. These drugs have significant side-effects [39] and their efficacy is controversial as there is currently no way to measure cure rates without decades of follow-up. There is consequently a lack of consensus on treatment duration throughout the community studying Chagas disease [40], and policy decisions and recommendations cannot be made. There is an urgent need for a reliable test that can determine cure in treated patients and measure treatment efficacy for a drug [41].

In our previous study we assessed cure in a cohort of adult patients treated with Nfx and followed up after three years by various diagnostic tools. Our results demonstrated the inadequacy of conventional serological tests for assessing cure in patients. In the cohort, 97.3% had results that could either indicate treatment failure or persistent humoral response despite having undergone treatment. On the other hand, the same samples all tested negative by conventional PCR and all but one tested negative by RT-PCR. These results reinforced the need for a robust and rapid test to assess treatment efficacy with certainty.

Research on the discovery and use of new antigens for diagnosis is on-going [42–44], but new diagnostic techniques such as proteomics [36] and flow cytometry [45] are also being investigated. In this study, we looked at the effect of Nfx in Chagas patients by using proteomics on samples from the same cohort of adults described in our previous study. We targeted our research on key biomarkers previously identified in our laboratories for diagnostic purpose [36] with a two-fold objective: assess whether specific biomarkers could be discriminative enough and help in defining cure in patients on the one hand and could be used as potential antigens for further development of new test of cure on the other hand.

Using SELDI, we successfully identified the majority of the key biomarkers discovered in our previous study [36] in the three test groups (healthy, infected and Nfx-treated follow-up samples), focusing on key biomarkers (mature human APOA1 (28.1 kDa); fragments of human APOA1 (24.7, 13.6 and 9.3 kDa); and a fragment of human FN1 (28.9 kDa). The pattern of these biomarkers between healthy and diseased samples was consistent with those seen in our previous study, i.e. all truncated fragments were up-regulated during disease compared to healthy controls whereas full-length APOA1 was downregulated during disease compared to healthy controls. These results, based on samples from Bolivian patients, confirm those previously seen with Venezuelan patient samples and demonstrate the reproducibility of SELDI technology across populations from South America [36]. Cluster plots showed that the levels of these biomarkers in follow-up samples of patients treated with Nfx returned to levels similar to those seen in healthy controls.

Seroconversion is currently the only surrogate marker of cure for Chagas disease. Given that all patients in this cohort were seropositive at the three year follow-up, we could not make any correlation with seroconversion.

Our five chosen biomarkers were shown to achieve up to 100% sensitivity and 98% specificity for CD, enabling the classification of follow-up samples as coming from healthy (or cured) or Chagas subjects. The results indicated that Nfx had no effect in 56.2% (n = 21/37) of the patients, one of whom was known to be still infected with T. cruzi (positive RT-PCR).

Nfx treatment length was found to be important, as nine out ten patients who pursued treatment for less than 40 days were still classified as Chagas disease patients. Of the 27 patients who followed Nfx treatment for more than 40 days, 25 of whom complied with the full 60 days of treatment, 15 (55.6%) were classified as healthy and would be considered as cured.

Taken together these results suggest that these biomarkers might be useful in assessing treatment efficacy in chagasic patients and could lead to the development of a test of cure. Moreover, their evaluation in animal models of Chagas disease could lead to improvements in the predictability of these models and their translation to the human disease during the drug discovery and development process.

Although very encouraging, it is important to point out that our study has limitations, and further studies will be needed to establish the real potential of these markers for their use as test of cure and assessment of treatment efficacy for Chagas disease. We were unable to correlate the biomarkers data with seroconversion, the only surrogate for Chagas disease, and had to rely on the discriminatory power of an algorithm using these markers. Since seroconversion in children is much quicker than in adults (1–2 years and decade(s) respectively), a follow-up study with sera samples from treated children would confirm a correlation between these biomarkers and seroconversion, and therefore cure. Indeed, access to samples of adults treated for Chagas disease with follow-up samples for decades is very limited. We also did not have serum samples corresponding to End of Treatment (EoT) for analysis. Indeed, being able to establish cure and perform counseling at the end of treatment would already be a major breakthrough, and would eliminate the need for a long follow-up. In addition, it may be possible to establish if down-regulation of key biomarkers is directly linked to the absence of cruzipain (cruzipain cleavage sites have been identified in the APOA1 and FN1, data not shown).

Finally, access to and analysis of samples originating from larger cohorts of patients treated with other drugs (e.g. Benznidazole) would be very useful and could further validate our findings.

There is a lack of reliable tests for assessment of cure following treatment in chronic Chagas disease patients. In this study we identified a biomarker pattern strongly associated with CD. Using an algorithm, we showed that Nfx-treated patients that have the same biomarker profile/pattern as healthy controls were classified as cured. Further studies using samples from chagasic children would be useful to address the correlation between these biomarkers and seroconversion.