Rapid Discrimination of Malaria- and Dengue-Infected Patients Sera

Subscriber access provided by OCCIDENTAL COLL

Article

Rapid discrimination of Malaria and Dengue Infected Patients Sera using Raman Spectroscopy Sandip Kumar Patel, Nishant Rajora, Saurabh Kumar, Aditi Sahu, Sanjay K Kochar, C. Murali Krishna, and Sanjeeva Srivastava Anal. Chem., Just Accepted Manuscript • Publication Date (Web): 29 Apr 2019 Downloaded from http://pubs.acs.org on April 29, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Rapid Discrimination of Malaria and Dengue Infected Patients Sera using Raman Spectroscopy

Sandip K. Patel1, Nishant Rajora1, Saurabh Kumar1, Aditi Sahu2, Sanjay K. Kochar3, C. Murali Krishna2*, Sanjeeva Srivastava1* 1 Department

of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India 2

Chilakapati lab, ACTREC, Tata Memorial Center, Kharghar, Navi Mumbai-410210, India

3 Department

of Medicine, Malaria Research Center, S.P. Medical College, Bikaner 334003, India

*Correspondence:

1. Professor Sanjeeva Srivastava, Department of Biosciences and Bioengineering, Indian Institute of Technology Bombay, Powai, Mumbai 400076, India E-mail: [email protected]; Phone: +91-22-2576-7779, Fax: +91-22-2572-3480 2. Professor Chilakapati Murali Krishna, Chilakapati lab, ACTREC, Tata Memorial Center, Kharghar, Navi Mumbai-410210, India E-mail: [email protected], [email protected]

1 ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 18

Abstract Malaria and dengue have overlapping clinical symptoms and prevalent in the same geographic region (tropical and sub-tropical), hence precise diagnosis is challenging. High mortality-rate associate with both malaria and dengue could be attributed to “false”, “delayed” or “missed” diagnosis. The present study thus aims to stratify malaria and dengue using Raman Spectroscopy (RS). In total 130 human sera was analyzed for model development and double-blinded testing. Principal Components-Linear Discriminant Analysis (PC-LDA) of acquired RS-spectra could classify malaria and dengue with a minor overlap of 16.7%. Receiver Operating Characteristic (ROC) analysis of test samples showed sensitivity/ specificity of 0.9529 for malaria vs Healthy Control (HC) and 0.9584 for dengue vs HC. The Raman findings were complemented by Mass Spectroscopy (MS) based metabolite analysis of 8 individuals each from malaria, dengue and HC. Several of the metabolites including amino acids, cell-free DNA, creatinine and bilirubin assigned for the predominant RS-bands were also identified by MS and showed similar trends. Our data clearly indicates that RS-based serum analysis using microprobe has immense potential for early, accurate and automated detection and discrimination of malaria and dengue, and in future could be extrapolated in field-settings combined with handheld RS. Further, this approach might be extended to diagnose other closely related infections with similar clinical manifestations. Keyword: Malaria, Dengue, Raman Spectroscopy, Mass Spectrometry, Metabolomics, Serum Introduction Malaria remains one of the most devastating global health problems affecting approximately 40% of the world’s population. Every year, malaria affects more than 500 million people and a child dies every 30 seconds 1. The good news, however, is that in most WHO regions, the proportion of suspected malaria cases receiving parasitological test among patients presenting for treatment in the public sector has increased. And the largest increase has been in the WHO-African Region, where diagnostic testing increased from 40% of suspected malaria cases in 2010 to 76% in 2015. However, 24% of cases do not even undergo diagnostic testing. Present malaria diagnosis relies on the identification of Plasmodium, causative agent of malaria in blood and the “gold standard” being histopathological analysis of Giemsa stained blood-smears. However, the results are subjective and accurate identification requires experienced technicians. Many-a-times asymptomatic malaria or that with low-parasitemia levels goes undetected 2. Numerous alternative approaches including Rapid diagnostic testing (RDT), Polymerase chain reaction (PCR), serological tests and mass spectrometry (MS) have been explored to develop diagnostic tools that are more reliable than the current gold standard but often fails. WHO recommends malaria RDT (MRDT) diagnosis for all people with suspected malaria before treatment is administered. MRDT kit detects persisting antigens such as histidine-rich protein II (HRPII) and lactate dehydrogenase for falciparum and vivax infection 2 ACS Paragon Plus Environment

Page 3 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


respectively. However, the antigen remains in circulation for a while once it reaches peak parasite density, thus leading to high frequencies of “false positive” HRPII detection3. Besides “false negative” HRPII detection was reported with hrp2-deleted falciparum isolates 4. Based on these observations, the Centers for Disease Control and Prevention (CDC) recommended that MRDT does not eliminate the need for malaria microscopy and all negative and positive cases must be followed by microscopy. Unfortunately, many areas of endemicity lack resources and due to remoteness has limited microscopybased treatment monitoring. A recent study compared the efficacy of microscopy, RDT and PCR based malaria diagnosis and confirmed the superiority of PCR over the other two methods, nevertheless authors questioned its applicability for routine diagnosis or field-settings based on cost, timeconsumption, prerequisite of trained personnel, cold-chain reagents, and sample processing (gDNA extraction and gel run)5. Studies show that missed or delayed diagnosis of malaria in 59-71% of imported cases 6,7 or in non-endemic regions, for instance, dengue endemic area 8,9 adds-on to morbid and fatal outcomes. Global escalation of dengue incidence, approximately 3.9 billion cases in 128 countries have worsened the situation 10. CDC suggests that the earliest diagnosis of acute dengue infection can be established by testing sera during the first 5 days of symptoms and/or early convalescent phase, contrary to the fact, the rapid and accurate the diagnosis is, the better the patient management and recovery due to lack of specific therapy and vaccine against dengue 11. Dengue diagnosis relies on serological tests (primarily ELISA for detection of IgM and IgG antibodies), dengue virus isolation, and molecular methods (detection of viral nucleic acid/antigen). Among these, the most commonly used combination includes detection of viral antigen, non-structural protein 1 (NS1) and antibody, IgM. A data report from Martinique during co-epidemic of dengue virus serotypes 2 and serotype 4 demonstrated that only 67.1% of patient sera showed positive for dengue NS1 during early diagnosis

12.

Secondly, a false

positive diagnosis of dengue due to cross-reactivity of antibody responses with other circulating flaviviruses has been a confounding issue in providing a differential diagnosis13-15. Interestingly, a report on the evaluation of ELISA-based serodiagnosis of dengue fever in travelers highlighted that IgM response could be detected only 4-8 days after onset of the clinical symptoms16. The down-side of IgMELISA being the requirement of acute sample, expertise and appropriate facilities, time-consuming (sometimes more than 1 week), expensive thus screening entire-set of suspected patients having fever requires huge cost, cannot differentiate between primary and secondary infection, IgM levels at times are below recognition due to secondary infections

17

and may lead to misdiagnosis

physicians in endemic areas sometimes only use positive tourniquet test

19

17,18.

Thirdly,

and start treatment which

might lead to unnecessary treatment of false-positive cases18,20. Conversely, in geographic locations where dengue is less frequent, problems of missed-diagnosis are often apparent8. Dengue infections with atypical clinical appearances such as hepato-gastrointestinal, neurological, and cardiac presentation remain underdiagnosed. Fourthly, a recent survey conducted by CDC and the Angola 3 ACS Paragon Plus Environment


Page 4 of 18

Ministry of Health is a whistle-blower, where patients in the endemic regions of Africa confirmed as positive using dengue RDT kit including one patient with symptoms consistent with severe dengue turned-out to be false-positive21. A recent article in the nature review microbiology also highlighted the limitations of current dengue diagnostic tests17. Simply relying on detection of particular antigen or antibody is not sufficient, and sometimes misleading for dengue diagnosis, particularly in case of concomitant infections with malaria16. Lahiri et al. reported that two-thirds of deaths from dengue occurred due to missed diagnoses16. Considering the diagnostic limitations, this study aims to develop diagnostic-methods for simultaneous discrimination of malaria and dengue infection using RS. High chemical specificity, low operational cost, minimal or lack of sample preparation and and integrated machine learning to automate Yes/No response have recently led to an increased application of RS in medical diagnostics. Timecourse study of mice plasma infected with Plasmodium clearly indicated that RS could accurately detect early stages of infection, with parasitemia levels as low as 0.2%, which is typically difficult to be detected by existing methods 22. Recently handheld NanoRam® spectrometer has been used to evaluate the quality of anti-malarial drugs23. Likewise, RS has also been used to analyze dengue fever in infected patient sera24-26. Similarly, MS-based metabolic-profile shift during intra-erythrocytic development of P. falciparum has been shown27. Alterations in several metabolic pathways using global metabolic study of RBCs infected with P. falciparum have been reported 28. Besides, serum proteomic profiling for diagnosis and progression of malaria and dengue have been performed independently by our group29-31. Correspondingly, in the present study 130 human sera were included for RS model development, while 42 samples were used for double-blinded testing of the model. Experimental Ethics statement Approval was taken from the Institutional Ethics Committees of Sardar Medical College, Bikaner, Rajasthan India. Experiments were performed in accordance with guidelines and regulations for the involvement of human subjects. Subject recruitment and sample collection Blood samples were collected from 37 malaria and 39 dengue patients, and 54 HC (patient inclusionexclusion criteria were used). Malaria cases were confirmed by the thick blood smear microscopy method and RDT, while dengue infections were diagnosed by IgM antibody capture enzyme-linked immunosorbent assay. Clinical details of the patients are enlisted in Table S1. Plasma separation was performed as per the previously described protocol 29-31. Raman spectroscopy Frozen serum (30 μl) of malaria, dengue, and HC following passive thawing was individually placed on CaF2 window to record spectra using HE-785 commercial RS (Jobin-Yvon-Horiba, France) with fiber-optic microprobe following the previously described protocol 4 ACS Paragon Plus Environment

32.

Spectra were acquired in the

Page 5 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


range of 500-3500 cm-1 for 20 seconds and averaged over three accumulations; acquisition details: laser power = 30 mW and excitation wavelength (λex) = 785 nm. Eight spectra were recorded for each sample and randomized across different sample classes. Data analysis was performed using MATLAB (Mathworks Inc.) based in-house software. A detailed protocol is provided in SI Materials and Methods. Mass spectrometry Serum metabolites were extracted using methanol 33, data acquisition was done by Agilent 1260 Infinity HPLC system coupled to an Agilent 6550 HRLC system, and analyzed by MassHunter™ Qualitative Analysis B.05.01 (Agilent) software. A detailed protocol is provided in SI Materials and Methods. Absolute quantification of creatinine and bilirubin Creatinine and bilirubin were quantified using automated blood analyzer (Transaria Bio-medicals LTD in collaboration with ERBA Diagnostics Mannheim GmbH). Results and Discussion Clinicopathological analysis of 130 subjects, 37 malaria (31 male and 6 female), 39 dengue (32 male and 7 female) and 54 HC (42 male and 12 female), confirmed with RDT and microscopy for malaria and ELISA for dengue, indicated altered hematological and biochemical parameters in malaria and dengue patients (Table S1). Non-parametric Mann-Whitney U test was implemented to evaluate the statistical significance. Total bilirubin, urea, liver enzymes such as alanine aminotransferase (ALT), aspartate transaminase (AST), and alkaline phosphatase (ALP) were significantly elevated in both diseased groups compared to HC (Table S1), suggesting discrimination of malaria and dengue is difficult based on the clinicopathological features, therefore to overcome the limitations, in the present study we employed serum RS. Besides, it is apparent that malaria and dengue cases in males outreached female patients; hence we tried to maintain a similar trend in HC to eliminate any biasness that might influence statistical analysis. There are several research papers published on malaria instances and associated-death from certain hypoendemic parts of the world 21 and India 34-36, which strongly support the fact that males (age groups—18-56) are more susceptible to malaria as compared to females. However, any co-relation between sex hormones and malarial susceptibility in human remains ambiguous37. Similar trends in demographic factors such as age and sex were mirrored in dengue infections38,39. Raman spectra for HC (n=34), malaria (n=27) and dengue (n=29) sera were recorded to generate a predictive model using compact and portable microprobe. Mean of eight Raman spectra for malaria vs HC (Fig. 1A i) and dengue vs HC (Fig. 1B i) is presented. Major spectral peaks were identified for β-carotene (1157 cm-1), amide linkages (1280, 1302, 1337 cm-1), CH2 deformation (1337, 1398, 1445 cm-1), cell-free DNA (1340, 1420 cm-1), Tyr (830, 850 cm-1), Trp (1552 cm-1), Phe (1004, 1204 cm-1), Creatine (846, 908 cm-1), Bilirubin (950, 968 cm-1), Protoporphyrin IX (1619, 1585, 1339, 1255, 970 cm-1), Cytochrome-like moiety (1553 cm-1), Asn (1594 cm-1), Glu (1537, 1075, 1318 cm-1), Pro (1045 cm-1), Galactosamine (1514, 1387, 1318 cm-1), Triglycerides (1075, 1300 cm-1), Cellobiose 5 ACS Paragon Plus Environment


(885 cm-1), and Glucose (1117 cm-1). Interpolated first derivative area-normalized spectra were used for multivariate analysis to classify samples using supervised Principal components-linear discriminant analysis (PC-LDA) methods, and validation using Leave-one-out cross-validation (LOOCV) respectively. Firstly, 2-model system (malaria vs HC, Fig. 1A and dengue vs HC, Fig. 1B) was employed; subsequently, 3-model system (malaria vs dengue vs HC, Fig. 1C) was developed for more robust testing of our data-sets. Once the predictive model was generated, double-blinded samples comprising of 20 HC, 12 malaria and 12 dengue sera was evaluated for the validation-of-proof (Table 1). Factor 4 was used for the analysis of malaria vs HC. A scree plot, depicting variance or percent correct classifications, accounting for approximately 90% correct classifications in case of malaria vs HC is shown in Fig. 1A ii. Likewise, factors 3 was used for PC-LDA analysis of dengue vs HC and the respective scree plot accounting approximately 95 correct classifications is shown in Fig. 1B ii. The PC-LDA scatter plot thus obtained for malaria vs HC (Fig. 1A iii) showed two minimally overlapping clusters, while the counterpart for dengue vs HC showed two clusters with no overlap (Fig. 1B iii). The results are also summarized in Table 1A for malaria vs HC and Table 1B for dengue vs HC respectively. As seen, 23/27 malaria-spectra and 32/34 HC-spectra were correctly classified. LOOCV analysis showed that two HC and four malaria spectra were misclassified (Table 1A), thus malaria-spectra could be classified with 85% efficacy. Likewise, LOOCV to evaluate the classification efficiency of the dengue vs HC model indicated an efficiency of 100% for dengue and 88% for HC (Table 1B). ROC showed sensitivity/specificity of 0.9529 for malaria vs HC (Fig. 1A iv) and 0.9584 for dengue vs HC (Fig. 1B iv). Factor 2 was used for PC-LDA analysis of three-model viz. malaria vs dengue vs HC. A scree plot accounting for approximately 85% correct classification is depicted in Fig. 1C i. The PCLDA scatter plot (Fig. 1C ii) indicated three clusters with minor overlap; the results are summarized in Table 1C. As indicated 25/29 dengue-spectra, 21/27 malaria-spectra and 29/34 HC-spectra were correctly classified. An average classification efficiency of 86%, 77%, and 85% was observed for dengue, malaria, and HC respectively. LOOCV evaluation showed classification efficiency of 83.3% for dengue and malaria respectively (Table 1C), while 100% efficacy was noted for HC-spectra. The overall precision of the model was 0.9-1.0. To the best of our knowledge, the classification efficiency that we achieved to stratify dengue and malaria is better than the existing diagnostic approaches. In the future, extensive studies employing larger-cohorts in combination with machine learning approaches could further improve classification efficiency. The model presented here is a proof-of-concept and will be challenged with different viral-infected sera and verified before clinical translations. Raman spectral interpretation can be challenging because of the complexity of biological samples. The aromatic amino acids- Phe, Tyr, and Trp are usually detected in Raman spectra and each has multiple peaks. Therefore, spectral annotation/assessment is done based on the presence of multiple peaks for a suspected biomolecule in accordance to the available literatures40-42. Moreover, we confirmed the robustness of our Raman model/data by complementing with LC-MS based global metabolomics study. In total 676 metabolites (p ≥ 0.05, fold change ≤ 2) were differentially regulated 6 ACS Paragon Plus Environment

Page 6 of 18

Page 7 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


in malaria, dengue and HC (n = 8 each) (Table S2). Volcano plot showing p-values versus log 2 metabolites ratio of malaria/HC (Fig. 2A i), dengue/HC (Fig. 2A ii), and dengue/malaria (Fig. 2A iii) are presented. 3D-PCA plot of the significantly altered features showed clear-cut clustering of malaria, dengue, and HC (Fig. 2B). Hierarchical clustering analysis (HCA) was performed on the combined sample classification with metabolites clustering to identify the metabolites important for patient sample grouping (Fig. 2C). Amino acids, glucose, creatinine, and bilirubin identified by MS corresponded to the Raman bands/ assigned molecules and severed as the basis for the discrimination of malaria and dengue (Table 2). Raman spectra peaks at 1154 and 1524 cm-1, assigned as β-carotene were observed in all sera samples with the highest relative intensity in HC, followed by malaria and dengue. Our findings are in corroboration with the previous report wherein the level of β-carotene in blood was found to be decreased in malaria

41.

Strong DNA peak (̴ 1344 cm-1) was observed in malaria sera,

implying parasite DNA and/or the lysed host cell DNA in circulation 40. Cytochrome-like moiety was found to be down-regulated in malaria, while the converse was observed in dengue, these finding are well justified by heme-degradation in malaria. Besides several of the amino acids including asparagine (Asn), glutamate (Glu), proline (Pro), phenylalanine, tryptophan and tyrosine was differentially modulated (Table 2). Plasmodium genome encodes a few enzymes synthesizing glycine (gly), proline (Pro), glutamate (Glu), glutamine, aspartate and asparagine (Asn) 43,44. Asn serves as one of the most abundant amino acids in Plasmodium proteins

45.

Deletion of asparagine synthetase in P. berghei delays the

asexual- and liver-stage development 44. Likewise, depletion of blood Asn levels in mice completely prevents the development of liver and sexual stages. Up-regulation of Asn in malaria as detected by RS may thus be used as a signature for malaria infection and in future could be targeted to prevent malaria transmission and liver infections. Asn is linked to dengue viral morphogenesis and infectivity46. Prorich domain of the dengue virus (DENV) is crucial for infectious particle production47, hence Pro was found to be up-regulated in dengue-infected patient sera. Conversely, the reduced levels of Pro in malaria could be associated with its accumulation in the food vacuole (up to 30-fold), often linked to antimalarial resistance in P. falciparum 48. Glu was down-regulated in both malaria and dengue; possible due to the enhanced energy needs the Glu is channeled to glycolysis 49. Likewise, glucose was downregulated in both malaria and dengue. DENV infection is associated with enhanced glucose consumption and depriving DENV-infected cells of exogenous glucose demonstrated a pronounced impact on viral replication50. Likewise, glucose consumption may increase up to 100-fold in Plasmodium-infected RBC

51.

Triglycerides were down-regulated in both malaria and dengue as

compared to HC 52. Additionally enhanced creatinine (Fig 2D) and bilirubin (Fig 2E) levels were noted in malaria and dengue patient sera. This study provides a comprehensive comparative representation of the Raman spectra to distinguish infectious disease (malaria and dengue) with overlapping clinical symptoms. There were several sections in the recorded spectra for healthy and diseased conditions that were overlapping. 7 ACS Paragon Plus Environment


However, the multivariate models that allow quantitative and objective diagnosis for independent patients are generated based on differential spectral patterning. The difference in spectra provides quantitative information regarding modulation in all the chemical constituents of biofluids including amino acids and metabolites enlisted in Table 2, as compared to the control spectra that might either be caused or are the cause of disease; thereby, classifying patients with different pathogenic infections. More importantly, few of the metabolites including Asn, Glu, glucose, triglycerides, bilirubin and protoporphyrin IX identified in RS showed similar trends in MS analysis. RS is more robust than RDTs or PCR that employs only one or two specific antigens/antibodies for disease diagnosis. Besides, RS has been successfully employed in diagnostics, prognostics or as a tool for evaluating new therapies for several diseases including cancer 32,53-55, urology and andrology 56, asthma 57, tuberculosis 58 and bone demineralization 59. Interestingly, our group was able to classify the serum samples of buccal mucosa and tongue cancers, two very closely related cancer using RS 57. The further in-depth study may lead to the identification of a novel candidate for drug targets. One of the drawbacks of the study is small sample size; however, we tried our best to compensate for this limitation using complementary MS-based approach to validate our findings. The RS model developed in the study is very reliable, accurate, time-efficient and requires minimal sample-processing. This is the first report on the systematic serum-based RS to distinguish malaria and dengue. Miniature versions including microprobe, used in the present study, and handheld devices could provide flexibility to be used in both laboratory- and field-settings. Comparisons of the sensitivity of field-portable and benchtop RS for the assessment of Point-of-care infectious disease diagnostics displayed promising results58. Likewise, the handheld Raman was used efficiently for the detection and identification of microbial pathogens isolated from human serum 60, diabetes 61, multiple myeloma 62, anti-malarial drugs quality 23, quantitative tuberculosis58 and plant pathogens on maize kernels in field-setting 63. We are very optimistic that in future the study could be translated to the field-settings particularly in malaria and dengue-endemic regions to screen larger patient cohorts.

Associated content Supporting Information is available free of charge via the Internet on the ACS Publications website http://pubs.acs.org Detailed Materials and Methods and Supplementary Tables including Table S1 (hematological and biochemical parameters of the dengue and malaria patient and healthy control) and Table S2 (Comprehensive list of all statistically significant (p < 0.05 and fold change 0.05) (B) 3D-PCA plot of significantly altered metabolites (p-value > 0.05 fold change >2). (C) Hierarchical clustering of significantly altered metabolites (p-value > 0.05 fold change >2). (D, E) Absolute quantification of creatinine and bilirubin respectively. ** indicate p < 0.001 and * indicate 0.001 < p < 0.05 based on Mann-Whitney test.


Page 12 of 18

Page 13 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 1: PC-LDA confusion matrix A Malaria vs HC B. Dengue vs HC C Malaria vs Dengue vs HC. (a) The standard model, (b) Leave-one-out cross-validation; wherein diagonal elements are true positive predictions and ex-diagonal elements represent false positive predictions and (c) Validation of the proof using blinded/test samples.

True condition

Malaria Healthy

True condition

Malaria Healthy

True condition

Dengue Healthy

True condition

Dengue Healthy

True condition

Dengue Malaria Healthy

True condition


True condition


A. Malaria vs Healthy (a) Standard model (PC-LDA) Predictive condition Malaria Healthy Total % correct decision 23 4 27 85.1 2 32 34 94.1 (b) Leave-one-out cross-validation (LOOCV) Predictive condition Malaria Healthy Total % correct decision 23 4 27 85.1 2 32 34 94.1 B. Dengue vs Healthy (a) Standard model (PC-LDA) Predictive condition Dengue Healthy Total % correct decision 29 0 29 100 4 30 34 88.2 (b) Leave-one-out cross-validation (LOOCV) Predictive condition Dengue Healthy Total % correct decision 28 1 27 96.5 4 30 34 88.2 C. Malaria vs Dengue vs Healthy (a) Standard model (PC-LDA) Predictive condition Dengue Malaria Healthy Total % correct decision 25 1 3 29 86.2 2 21 4 27 85.3 3 2 29 34 85.3 (b) Leave-one-out cross-validation (LOOCV) Predictive condition Dengue Malaria Healthy Total % correct decision 25 1 3 29 86.2 4 19 4 27 70.3 3 3 28 34 82.3 (c) Test data set Dengue Malaria Healthy Total % correct decision 10 1 1 12 83.3 0 10 2 12 83.3 0 0 20 20 100

13

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46

Page 14 of 18

Table 2: Selected list of entities modulated differentially in malaria and dengue infections as identified by Raman spectroscopy and mass spectrometry. Raman peak

Designated metabolites

Regulation in Dengue

Regulation in Malaria

Metabolites identified in MS

1594

Asparagine (Asn)

Up

Up

Identified

1537, 1075, 1318

Glutamate (Glu)

Down

Down

Other derivative

1045

Proline (Pro)

Up

Down

Non identified

16101616

Phenylalanine, tryptophan, and tyrosine

Up

Up

Other derivative

1075, 1300

Triglycerides

Down

Down

Other derivative

885

Cellobiose

Down

Not regulated

Non identified

1117

Glucose

Down

Down

Identified

846; 908

Creatinine*

Up

Up

Non identified

Putative roles Asn plays a vital role in DENV propagation. Plasmodium is strikingly rich in Asn. In fact plasmodium retains a gene encoding asparagine synthetase DENV induces and requires glycolysis for optimal replication, thereby channeling glutamate to the TCA cycle. Putative glutamine synthetase gene is found in Plasmodium. Besides plasmodium generates large amounts of ammonia (preferably by glutamate pathway) in the absence of intrinsic parasite detoxification mechanisms. Pro-rich domain of the DENV is crucial for infectious particle production. Malaria-iRBC and free parasites have limited capabilities for the biosynthesis of amino acids, thus uptakes Pro from plasma or host-cell hemoglobin. Abnormal amino acid metabolism may serve as an important factor in malaria and dengue pathogenesis Galactosamine is associated with IgM, which show positive correlation with the triglycerides. Identified in RS studies of dengue but its role in infection unknown Glucose consumption is increased during DENV infection and depriving DENV-infected cells of exogenous glucose had a pronounced impact on viral replication. Decreased glucose level in malaria is ascribed to pronouncedly increased glucose use and impaired glucose production caused by the inhibition of gluconeogenesis. Creatinine is up-regulated in malaria and dengue infection


Ref

25,26,44

50,64

47,65

66,67

25,26

25,26

50,68

Page 15 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46


950; 968 1619; 1585; 1339 1553 1154, 1524

Bilirubin*

Up

Up

Identified

Bilirubin is up-regulated in malaria and dengue infection

Protoporphyrin IX

Up

Up

Identified

Protoporphyrin IX shows positive corelation with bilirubin

Cytochromelike moiety

Up

Down

N/A

Down

Down

N/A

β-carotene

cytochrome-like moiety down-regulation in malaria is well justified by heme-degradation. Down-regulated durin infection


41


Figure 1 178x269mm (300 x 300 DPI)


Page 16 of 18

Page 17 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2 185x217mm (150 x 150 DPI)



82x44mm (300 x 300 DPI)


Page 18 of 18

Rapid Discrimination of Malaria- and Dengue-Infected Patients Sera

Recommend Documents