Background

The control of malaria still continues to pose a problem in endemic countries, especially due to emergence and spread of drug resistant Plasmodium parasites across the globe [1]. Plasmodium falciparum and Plasmodium vivax are responsible for the bulk of global morbidity and mortality statistics [1, 2]. The most dangerous human malaria species is P. falciparum with high risk of severe complications if not treated promptly [3]. Plasmodium vivax is the most geographically distributed species, mainly encountered in South East Asia (SEA), Latin Americas, and the horn of Africa. India accounted for ~ 79% of cases and ~ 83% of deaths reported in SEA in 2021 [1, 4]. Recently, there are increasing reports on the ability of P. vivax parasites to induce severe malaria and deaths in some patients [5,6,7].

Plasmodium falciparum parasites have developed resistance phenotypes to all major anti-malarial drugs, such as chloroquine (CQ), mefloquine (MQ), sulfamides, artemisinin (ART) and its derivatives (Fig. 1) [8, 9]. In contrast, some studies evidenced the emergence of P. vivax parasites resistant to fewer drugs, such as CQ and MQ [10, 11]. The current first-line uncomplicated P. falciparum malaria treatment policies rely on artemisinin-based combination therapy (ACT) which consist of a combination drugs (i.e., ART derivatives) with slow-acting drug referred to as partner drug [12]. Currently, six artemisinin-based combinations are recommended by the World Health Organization (WHO) for treating uncomplicated P. falciparum malaria viz. artesunate + sulfadoxine–pyrimethamine (AS + SP), artesunate + amodiaquine (AS + AQ), artesunate + pyronaridine (AS + PY), artemether + lumefantrine (AL), dihydroartemisinin + piperaquine (DHA + PPQ), and artesunate + mefloquine (AS + MQ) [1, 8]. CQ is commonly used for treating clinical P. vivax infections. The WHO recommended to associate CQ with primaquine (PQ), a hypnozoiticidal drug, given the ability of P. vivax parasites to induce relapses due to reactivation of dormant liver stages (i.e., hypnozoites). The association CQ + PQ, also known as radical cure, guarantees treatment of current infection and prevent recurrent infection due to relapses by killing blood and liver parasite stages [13, 14].

Fig. 1
figure 1

(Sources: [15, 25, 27,28,29,30,31,32])

Timeline of introduction and appearance of resistance to main anti-malarial drugs in the world and India. QN quinine, PPQ piperaquine, PQ primaquine, CQ chloroquine, PG proguanil, MQ mefloquine, ART artemisinin, SP sulfadoxine–pyrimethamine, AQ amodiaquine, ACT artemisinin based combination therapy, AS artesunate, AL artemether–lumefantrine, SLP sulfalene–pyrimethamine, CQR chloroquine-resistant, NE North east states

The emergence and spread of ACT-resistant P. falciparum parasites in SEA has profoundly thwarted malaria control and elimination objectives in the region [15, 16]. More recently, independent appearance of ACT-resistant P. falciparum parasites from two African countries Rwanda and Uganda [17,18,19]; has given basis to fear the spread of ACT-resistance in Africa in future. In this context, anti-malarial drug resistance surveillance is a key component to successful malaria control and elimination. Several tools have been developed for Plasmodium drug resistance and these consist of (i) in vivo drug efficacy studies, (ii) in vitro assessment of drug susceptibility in parasites, and (iii) tracking of molecular markers associated with drug resistance [20]. The latter tools are largely used for drug resistance surveillance given the high cost of in vivo studies and lack of standardization of in vitro studies [20]. Also, molecular tools are much easier to implement and/or scale-up, and give prevalence estimates of drug resistant Plasmodium parasites over time and space with possibility for within and between study comparisons [20].

In India, P. falciparum and P. vivax are the two predominant Plasmodium species with prevalence ratio varying between states, but overall ratio close to one [21,22,23]. Clinical spectrum of malaria infections with P. falciparum and P. vivax ranges from asymptomatic to severe malaria [3, 5]. India has continuously modified and adapted national drug policies since 1982 to efficiently and timely control anti-malarial drug resistance (Fig. 1). The current treatment policy endorses treatment of (i) uncomplicated P. falciparum malaria with either artesunate + lumefantrine (AL) in North Eastern states or artesunate + sulfadoxine + pyrimethamine (AS + SP) in rest of states, accompanied by single dose PQ, (ii) uncomplicated P. vivax malaria with CQ + PQ, and (iii) severe malaria with quinine, artemether, artesunate, or artemether for 48 h, followed by quinine or above mentioned a state-specific artemisinin-based combination [24]. In contrast, data on anti-malarial drug resistance in P. falciparum and P. vivax parasites are still insufficient and fragmented in India [8, 25, 26]. The present study describes the spatial and temporal distribution of main putative molecular markers associated with drug resistance in P. falciparum and P. vivax isolates collected between 1993 and 2019 among malarious patients from different areas of India.

Methods

Sample and diagnosis

This study has been approved by the institutional review board of ICMR-National Institute of Malaria Research (NIMR), India. Malaria samples were collected between 1993 and 2019 from different field sites and health facilities (hospitals and primary health centres) of different states and union territories (Additional file 1). On field sites, malaria infection was first detected using rapid diagnostic tests (RDT), and confirmed microscopically and by PCR assay. Two RDTs namely SD Bioline Pf/Pv® (Standard Diagnostics, Inc., South Korea) and FalciVax™ (Pf/Pv) (Zephyr Biomedical, Verna, Goa, India) were used in this study. SD Bioline Pf/Pv® is a P. falciparum and P. vivax detecting RDT which targets PfLDH + PvLDH antigens. It has a panel detection score of 99.0% and 97.1% at 200 parasites/µL for P. falciparum and P. vivax, respectively. FalciVax™ (Pf/Pv) is a P. falciparum and Pv/Pvom detecting RDT which targets PfHRP2 + PvLDH antigens, with a panel detection score of 95.0% and 100% at 200 parasites/µL for P. falciparum and P. vivax, respectively [33].

The study samples comprised of both; clinically asymptomatic and symptomatic individuals of all age living in rural, semi-urban and urban areas of India. A total of 16 areas belonging to 14 states and union territories were study sites in the present study (Fig. 2). Details on urbanization level and malaria endemicity of study areas are presented in Additional file 2. The plasmodial DNA was extracted using QIAamp® DNA Mini Kit (Qiagen, Valencia, California, USA) as per manufacturer’s instructions in a final elution volume of 70 µL buffer (10 mM Tris–HCl; 0.5 mM EDTA; pH 9.0), and then stored at − 20 °C until needed. Plasmodial speciation was performed using polymerase chain reaction (PCR) protocols targeting the 18S subunit ribosomal unit gene of P. falciparum and P. vivax (Table 1) [34]. DNA sample was mixed in 25 µL PCR reaction containing 12.5 µL of 2X Go Taq green master mix (Promega Corporation, USA), 1 µL of each primer (10 µM), 1–2 µL DNA template, and free-nuclease water Q.S. PCR amplicons were loaded on 2% agarose gel pre-stained with ethidium bromide at 72 V for 1 h, and then visualized using an ultraviolet trans-illuminator. P. falciparum and P. vivax infections were confirmed by the presence of PCR bands of 205 bp and 120 bp, respectively [34].

Fig. 2
figure 2

Map of India showing study areas where Plasmodium isolates were collected. Each pie chart represents the total number of isolates analysed. P. falciparum (blue) and P. vivax (red). The size of pie chart is proportional to sample size. The map depicted here is taken from official website of Ministry of External Affairs, Government of India (https://mea.gov.in/india-at-glance.htm, accessed 15/11/2021)

Table 1 Main mutations in Pf and Pv drug resistance genes analyzed in the study

Amplification of P. falciparum and P. vivax drug resistance genes

Single-step and nested PCR protocols were used to amplify five genes associated with drug resistance in P. falciparum parasites viz. chloroquine resistance transporter (pfcrt), dihydrofolate reductase (pfdhfr), dihydropteroate synthase (pfhdps), multidrug resistance protein 1 (pfmdr1), and Kelch protein (pfk13) [8]. Also, orthologues of these genes in P. vivax isolates were analysed viz pvcrt-o, pvdhfr, pvdhps, pvmdr1 and pvk12, using published and developed singe-step and nested PCR protocols (Additional file 3).

Sequencing and SNP analysis

The amplicons were purified using GeneJet purification kit (Thermofisher) and sequenced in both directions based on Sanger dideoxy method. Sequencing were performed on in-house ABI 3730XL DNA analyzer (Applied Biosystem) with BigDye Terminator v3.1 sequencing kit (Applied Biosystem). Nucleotide and deduced amino acid of gene sequences were aligned and compared with references by using CLUSTALW program of MEGA X [35].

Plasmodium falciparum and P. vivax sequences of drug resistance genes were analysed in comparison with those of reference strains. The reference P. falciparum strains accession numbers were PF3D7_0709000 for pfcrt, PF3D7_0417200 for pfdhfr, PF3D7_0810800 for pfdhps, PF3D7_0523000 for pfmdr1, and PF3D7_1343700 for pfk13. The reference P. vivax strains used were PVX_087980 for Pvcrt-o, PVX_089950 for pvdhfr, PVX_123230 for pvdhps, PVX_080100 for pvmdr-1, and PVX_083080 for pvkelch12. The phylogenetic relatedness of P. falciparum and P. vivax isolates was done through BLAST of drug resistance sequences. After sequence alignment nucleotide positions which displayed two peaks at one locus in chromatogram were noted as ‘‘mixed’’ and excluded from further analysis. Known point mutations in P. falciparum genes associated with anti-malarial drug resistance (CQ, SP, ART and its derivatives), and novel mutations were identified using their corresponding amino acids and haplotypes [8, 36]. Regarding P. vivax isolates, putative drug resistance-associated mutations were also investigated [37,38,39,40] (Table 1). The proportions of each validated mutation and putatively associated with anti-malarial drug resistance in P. falciparum and P. vivax isolates were calculated by areas and year of collection. Similar analysis was made for drug resistance genotypes and haplotypes.

Statistical analysis

Data were keyed, coded and verified for consistency in an Excel spreadsheet (Microsoft Office, USA), and then exported to GraphPad v8.0.2 for Windows (GraphPad PRISM, Inc., San Diego, CA, USA) and StatView v5.0 for Windows (SAS Institute, Inc., Chicago, USA) for statistical analysis. Data were summarized as percentages in tables and graphics. Changes in the prevalence of alleles and haplotypes over time were evaluated for statistical significance using Pearson’s independence χ2 statistics for trend. Statistical significance was set at p < 0.05.

Results

Plasmodium samples

A total of 593 samples were positive for 18S genes, and infections with P. falciparum and P. vivax were found as either mono-infection or mixed infection (Fig. 3a and Additional file 4). The highest proportion of mixed infections were seen in Karnataka (64.7%) and Madhya Pradesh (44.4%) (Fig. 3b). Only mono-infections with P. falciparum or P. vivax were included in the study. The final number of samples included for each drug resistance gene varied from 30 to 318 based exclusion criteria (positive PCRs, good PCR bands, successful purification, successful sequencing, good quality sequencing) (Additional file 4). Additional file 5 depicts gel electrophoresis of P. falciparum and P. vivax drug resistance gene PCR results.

Fig. 3
figure 3

Electrophoresis gel depicting P. falciparum mono-infection, P. vivax mono-infection and mixed infection (a), and proportion of types of infections with P. falciparum and/or P. vivax (b). In a, the gel image is showing the 18S RNA PCR amplification of Plasmodium species. Lane 1: Plasmodium falciparum (205 bp). Lane 2: Plasmodium vivax (120 bp). Lane 3: mixed infection (Both P. falciparum and P. vivax). Lane L: 100 bp Ladder. In b, the international codes of areas were used. AS Assam, CG Chhattisgarh, DL Delhi, GA Goa, HR Haryana, JH Jharkhand, KA Karnataka, MH Maharashtra, ML Meghalaya, MP Madhya Pradesh, MZ Mizoram, OR Orissa, RJ Rajasthan, UP Uttar Pradesh

Profiling of P. falciparum anti-malarial drug resistance genes

Pfcrt

A total of 47 samples from Delhi, Haryana, Madhya Pradesh, Maharashtra, Goa, and Uttar Pradesh regions were analysed for the pfcrt gene. The analysis revealed high proportions of 72S and 76T mutations and absence of 73K mutation in samples from different parts of the country. All parasites from Delhi, Uttar Pradesh, and Goa were carrying these two mutations. The pfcrt 74I and 75E mutations were only reported from Maharashtra in equal proportion (50% each) (Fig. 4a). Three types of pfcrt genotypes viz. C72V73M74N75T76 (single mutant), S72V73M74N75T76 (double), and C72V73I74E75T76 (triple mutant) linked to anti-malarial resistance were reported (Fig. 3b). The double mutant SVMNT was predominant in most of the areas with proportions ranging from 75 to 100% between 1994 and 2019, with the exception of Maharashtra where 50% of isolates collected in 2012 had triple mutations (i.e., CVIET) (Fig. 4b). No novel mutation and synonymous mutations were found in pfcrt analysed sequences.

Fig. 4
figure 4

Pfcrt genotypes by year and area. a Proportion of 72S, 73K, 74I, 75E and 76T mutations in pfcrt gene, b proportion of pfcrt haplotypes. Pf: P. falciparum; crt: chloroquine resistant transporter gene. The international codes of areas were used. DL Delhi, GA Goa, HR Haryana, MH Maharashtra, MP Madhya Pradesh, UP Uttar Pradesh. In a, sample size was different for each area. DL (n = 3), GA (n = 4), HR (n = 18), MH (n = 10), MP (n = 4), UP (n = 8). In b the wild type is depicted in green while mutants are depicted in red and pink. The map depicted here is taken from official website of Ministry of External Affairs, Government of India (https://mea.gov.in/india-at-glance.htm, accessed 15/11/2021)

Pfdhfr and pfdhps

In total, 239 and 318 samples were successfully sequenced for pfdhfr and pfdhps genes (Additional file 5). Samples were collected from different regions of the country viz: Assam, Chhattisgarh, Delhi, Goa, Haryana, Jharkhand, Karnataka, Maharashtra, Meghalaya, Mizoram, Orissa, Rajasthan, and Uttar Pradesh.

On analysis of the pfdhfr gene, four of the six validated SNPs (i.e., 51I, 59R, 108N and 164L) were found in most areas. The mutations 59R and 108N were more frequently found at high proportions as compared to 51I and 164L. The proportions of these mutations ranged from 0 to 65.4% for 59R, 8.3–72.7% for 108N, 0–23.1% for 51I, and 0–21.1% for 164L (Fig. 5a). A total of 15 pfdhfr resistance genotypes were found across the areas as single (16.3%), double (25.5%), and triple mutants (7.9%). Genotype richness was seen in Delhi (10 genotypes) and Haryana (12 genotypes) even though a high proportion of P. falciparum isolates were wild type (Fig. 5b). The single mutant A16C50N51C59N108I164 was mainly found in Uttar Pradesh at a proportion of 55.6%, while the double mutant A16C50N51R59N108I164 was mostly reported seen Raipur (57.8%) and Goa (60%). The triple mutant A16C50I51R59S108L164 was reported only in Haryana (1.9%) while the double mutant A16C50N51C59N108L164 was seen only in Jharkhand (4.5%) (Fig. 4b). Interestingly, the richness in resistance genotypes increased over years in Delhi with two, seven and eight genotypes in 1994, 2000 and 2008–2010, respectively. A specific distribution of pfdhfr haplotypes between two areas of Orissa (i.e., Rourkela and Bissam Cuttack—BCK) was noted. The mutants A16C50I51C59S108I164, A16C50N51C59N108I164 and A16C50N51R59N108I164 were found only in Rourkela while A16C50I51C59N108I164 and A16C50I51C59N108 L164 were found only in BCK (Fig. 5b). No novel mutation and synonymous mutations were found in pfdhfr and pfdhps sequences analysed.

Fig. 5
figure 5

Pfdhfr genotypes by year and area. a Proportion of pfdhfr 16 V, 50R, 51I, 59R, 108N and 164L mutations, b proportion of pfdhfr haplotypes. Pf: P. falciparum; dhfr: dihydrofolate reductase gene; wt: wild type. The international codes of areas were used. CG Chhattisgarh, DL Delhi, GA Goa, HR Haryana, JH Jharkhand, MH Maharashtra, MP Madhya Pradesh, OR Orissa, UP Uttar Pradesh. In a, sample size was different for each area. CG (n = 13 and 26), DL (n = 12, 17, and 37), GA (n = 5), HR (n = 53), JH (n = 22), MH (n = 9), OR (n = 17 and 19), UP (n = 9). In b wild alleles are in lower case and mutant alleles are in upper case. The map depicted here is taken from official website of Ministry of External Affairs, Government of India (https://mea.gov.in/india-at-glance.htm, accessed 15/11/2021)

For pfdhps, the mutations 436A/F, 437G and 540E were most frequently seen across areas as compared to 431V, 581G and 613S. The proportion of 436A/F mutation varied from 0% (in Goa, Madhya Pradesh, Rajasthan, Uttar Pradesh, and Mizoram) to 81.8% (in Chhattisgarh). All P. falciparum isolates from Goa, Karnataka, and Mizoram were carrying the 437G mutation while highest rates of 540E mutation were seen in Jharkhand (81.8%), Chhattisgarh (65.4%) and Meghalaya (50%) (Fig. 6a). Wild type-like haplotype accounted for 59.1% of all isolates, while the rest consisted of single mutants (11.9%), double mutants (22%) and triple mutants (6.9%). A higher genotype richness was found on analysis of pfdhps gene compared to pfhdfr gene, with 17 resistance genotypes. The highest number of genotypes were found in Delhi in 2000 (7 genotypes), Delhi in 2008/10 (10 genotypes), and Haryana in 2015/16 (11 genotypes) (Fig. 6b). The double mutant I431A436A437E540A581A613 was most spread as found in five areas at proportion of 14.2–21.6% (Delhi), 9.4% (Haryana), 72.8% (Jharkhand), 61.5% (Chhattisgarh), and 7.4–15.4% (Orissa). The wild type haplotype was found in 100% of P. falciparum isolates from Rajasthan, Madhya Pradesh, and Uttar Pradesh, while the double mutant I431S436G437K540G581A613 was found in 100% of isolates from Goa, Karnataka, and Mizoram (Fig. 5b). The double mutant I431S436A437E540G581A613 was found only in Delhi (4.8%) while triple mutants I431A436A437E540G581A613 and I431A436G437N540A581A613 were found only in Chhattisgarh (3.9%) and Assam (8.3%), respectively. Five of the nine isolates with triple mutant I431A436G437E540A581A613 were found in Meghalaya.

Fig. 6
figure 6figure 6

Pfdhps genotypes by year and area. Pf: P. falciparum; dhps: dihydropteroate synthase gene; wt: wild type. The international codes of areas were used. AS Assam, CG Chhattisgarh, DL Delhi, GA Goa, HR Haryana, JH Jharkhand, KA Karnataka, MH Maharashtra, ML Meghalaya, MP Madhya Pradesh, MZ Mizoram, OR Orissa, RJ Rajasthan, UP Uttar Pradesh. In a, sample size was different for each area. AS (n = 12), CG (n = 13 and 26), DL (n = 27, 21, and 37), GA (n = 8), HR (n = 53), JH (n = 22), KA (n = 5), MH (n = 10), ML (n = 12), MP (n = 6), MZ (n = 3), OR (n = 27 and 19), RJ (n = 10), UP (n = 6). a Proportion of pfdhps 431V, 436A/F, 437G, 540E/N, 581G and 613S mutations, b proportion of pfdhps haplotypes. In b wild alleles are in lower case and mutant alleles are in upper case. Samples from Gautaum Budh Nagar (UP) was excluded from percentage calculation because of low sample size (n = 1). The map depicted here is taken from official website of Ministry of External Affairs, Government of India (https://mea.gov.in/india-at-glance.htm, accessed 15/11/2021)

By combining SP resistance related pfhdfr and pfdhps mutations, we found 56 pfhdfr-pfhdps haplotypes represented by single mutants (six types), double mutants (15 types), triple mutants (12 types), quadruple mutants (14 types), quintuple mutants (6 types), and sextuple mutants (3 types) (Additional file 6). The quadruple mutant ACNRNI–IAAEAA accounted for 62.8% (27/43) of all quadruple mutants reported in the study, and was mainly seen in Chhattisgarh (Raipur). Two types of quintuple mutants viz. ACNRNL–IAAEAA and ACNRNI–IAAEAS accounted for 42.9% (6/14) and 28.6% (4/14) of all quintuple mutants, and were reported in Delhi, Orissa, and Jharkhand. To be noted, one isolate with quintuple mutant genotype (ACIRNI–ISGEAA) was found in Delhi. The sextuple mutants consisted of ACNRNL–IAAEAS (two isolates), ACICNL–IAAEAS (one isolate) and ACIRNI–IAAEAS (one isolate); and were found in Delhi, Orissa, and Haryana (Additional file 6).

Pfmdr1

Two of the five resistance pfmdr1 mutations (i.e., 86Y and 184F) were found in 135 samples from seven regions (Chhattisgarh, Delhi, Goa, Haryana, Jharkhand, Maharashtra, and Uttar Pradesh). The proportion of 86Y and 184F mutations ranged from 0 to 90.9% and 0 to 100%, respectively (Fig. 7a). Three types of mutants were found in this study, and were represented by Y86Y184S1034N1042D1246, N86F184S1034N1042D1246, and Y86F184S1034N1042D1246. The double mutant Y86F184S1034N1042D1246 was found only in Delhi (18.2%) and Uttar Pradesh (40%). Wild type-like pfmdr1 isolates were found only in Maharashtra (50%) (Fig. 7b). One synonymous pfmdr1 mutation (G182G) was found in 25 isolates from Goa (n = 4), Maharashtra (n = 2), Uttar Pradesh (n = 5), and Delhi (n = 14).

Fig. 7
figure 7

Pfmdr1 genotypes by year and area. a Proportion of pfmdr1 86Y, 184F, 1034C, 1042D and 1246Y mutations, b proportion of pfmdr1 haplotypes. Pf: P. falciparum; mdr1: multidrug resistance protein 1 gene; wt: wild type. The international codes of areas were used. CG Chhattisgarh, DL Delhi, GA Goa, HR Haryana, JH Jharkhand, MH Maharashtra, UP Uttar Pradesh. In a, sample size was different for each area. CG (n = 26), DL (n = 16), GA (n = 4), HR (n = 53), JH (n = 22), MH (n = 9), UP (n = 5). In b wild alleles are in lower case and mutant alleles are in upper case. The map depicted here is taken from official website of Ministry of External Affairs, Government of India (https://mea.gov.in/india-at-glance.htm, accessed 15/11/2021)

Pfk13

In this study, pfk13 sequences of 90 samples from Haryana (n = 53), Orissa (n = 8), Uttar Pradesh (n = 3), Karnataka (n = 4), Madhya Pradesh (n = 4) and Delhi (n = 18) were analysed. All isolates analysed for mutations in pfk13 gene were wild type, and no synonymous or nonsynonymous mutations for validated point mutations were observed. No novel mutation and synonymous mutations were found in pfk13 sequences analysed.

Profile of P. vivax drug resistance genes

Pvcrt-o

Thirty-two samples from Delhi were analysed in this section. Genetic profiling of Pvcrt sequences was available only for samples collected from Delhi in 2017. On analysis, the K10 “AAG” insertion was not detected in any sample.

Pvdhfr and pvdhps

In total, 117 and 126 samples from Delhi, Mewat, and Goa regions were analysed for drug resistance mutations in pvdhfr and pvdhps, respectively. Of the six pvdhfr point mutations analysed in sequences, amino acid changes were detected only in codons 58 and 117, with proportions ranging from 0 to 50% and 0 to 45%, respectively. The proportions of these two mutations were lower in Haryana compared to Delhi, and Goa (Fig. 8a). Similarly, amino changes were detected only in two of the nine pvdhps codons analysed namely 383 and 512. Also, all pvdhps mutations were found in P. vivax isolates from Goa, and were represented by 383G (59.1%) and 512N (13.6%) (Fig. 8b).

Fig. 8
figure 8figure 8

Pvdhfr and Pvdhps genotypes by year and area. a Proportion of pvdhfr 57L/I, 58R, 61M, 117N/T and 173F mutations, b proportion of pvdhps 382F/A/C, 383G, 399I, 512N, 585G, 553G, 555R, 585G, and 661V mutations, c proportion of the pvdhfr and pvdhps haplotypes. Pv: P. vivax; dhfr: dihydrofolate reductase; dhps: dihydropteroate synthase; wt: wild type. The international codes of areas were used. DL Delhi, GA Goa, HR Haryana. In a, sample size was different for each area. DL (n = 8, 32, 22, and 20 in years 2008, 2017, 2018 and 2019), GA (n = 22), HR (n = 13). In b, sample size was different for each area. DL (n = 28, 42, 22, and 12 in years 2008, 2017, 2018 and 2019), GA (n = 22). In c wild alleles are in lower case and mutant alleles are in upper case. The map depicted here is taken from official website of Ministry of External Affairs, Government of India (https://mea.gov.in/india-at-glance.htm, accessed 15/11/2021)

The genotype analysis revealed three and two mutated genotypes for pvdhfr (F57R58T61S117I173, F57S58T61N117I173 and F57R58T61N117I173) and pvdhps (S382G383M399K512R525A553K555V585A661 and S382G383M399N512R525A553K555V585A661), respectively (Fig. 8c). Wild type isolates accounted for 89.6% of all pvdhps genotypes.

The combination of pvdhfrpvdhps genotypes revealed nine haplotypes where most were represented by wild type. No mutant pvdhfrpvdhps haplotypes were found in P. vivax isolates from Delhi collected in 2017, but single mutant FSTNI–SAMKRAKVA (28.6%) and double mutant FRTNI-SAMKRAKVA (21.4%) were found in P. vivax isolates collected in 2018. In Goa, mutants were mainly represented by FSTSI–SGMKRAKVA (16.7%), FSTNI–SAMKRAKVA (13.3%) and FRTSI–SAMKRAKVA (10%) (Fig. 7c). No novel mutation and synonymous mutations were found in pvdhfr and pvdhps sequences analysed.

Pvmdr1

On analysis of 90 nucleotide sequences for pvmdr1 gene, amino acid changes were detected at all codons analysed with the exception of codon 976. The 845F, 861E and 898E were least frequently found mutations in P. vivax isolates with proportions of 0–13.3%, 0–3%, and 0–3%, respectively. The 908L mutation were at proportion of 39.4–100% in Delhi and 100% in Haryana. All P. vivax isolates carried 958M and 1076L mutations (Fig. 9a). No novel mutation and synonymous mutations were found in pvmdr1 sequences analysed.

Fig. 9
figure 9

Pvmdr1 genotypes by year and area. a Proportion of pvmdr1 845F, 861E, 898E, 908L, 958M, 976F and 1076L mutations, b proportion of pvmdr1 haplotypes. Pv: P. vivax; mdr1: multidrug resistance protein 1, wt: wild type. The international codes of areas were used. DL Delhi, HR Haryana. In a, sample size was different for each area. DL (n = 33, 18 and 24 in years 2017, 2018 and 2019), HR (n = 15). In b wild alleles are in lower case and mutant alleles are in upper case. The map depicted here is taken from official website of Ministry of External Affairs, Government of India (https://mea.gov.in/india-at-glance.htm, accessed 15/11/2021)

The genotype analysis pointed out the absence of wild type L845A861T898M908T958Y976F1076 in all areas. Five pvmdr1 genotypes were reported in this study, and were predominantly represented by triple mutant LATLMYL in isolates from Haryana (86.7%) and isolates from Delhi collected in 2018 (88.9%) and 2019 (100%). In contrast, double mutant LATMMYL accounted for 54.5% of all mutants found in isolates from Delhi collected in 2017. One quadruple mutant FATLMYL was found in Delhi—2018 (11.1%) and Haryana (13.3%) (Fig. 9b).

Pvk12

The analysis of 30 pvk12 sequences from Delhi and Goa regions revealed no mutations at codons 88, 124, 552, 581 and 697. Two novel mutations K264R (nonsynonymous) and L617L (synonymous) were found in Goa samples at proportions of in 27.3% (6/22) and 4.5% (1/22), respectively.

Discussion

The present study aimed at delineating genetic profile of main genes associated with drug resistance in P. falciparum and P. vivax malaria over 30 years in India.

Profile of anti-malarial drug resistance P. falciparum genes

Pfcrt

Most of pfcrt genotypes in study areas were double mutant SVMNT due to high rates of single mutations 72S and 76T, and this finding corroborates with previous reports [41,42,43]. In contrast, other studies reported lower rates of this double mutant in Odisha, and Arunachal Pradesh [44, 45]. A detailed analysis of the nucleotide codon at position 72 revealed that all SVMNT mutants from this study were of type S(agt)VMNT reported to have originated from Papua New Guinea [46]. The triple mutant CVIET was predominantly observed only in Maharashtra, which is not in line with previous reports from the same state [41]. The triple mutant was also reported from other states such as Odisha, Arunachal Pradesh, Chhattisgarh, and Assam [41, 43,44,45]. Area and time related drug policy changes could explain these between-study differences of SVMNT and CVIET proportions. The epidemiological profile of pfcrt genotypes hugely varies outside India, where triple mutant CVIET is predominant in countries, such as Cameroon and Saudi Arabia [47, 48], while wild type CVMNK is predominant in Ethiopia, Malawi, and Tanzania [49,50,51].

Pfdhfr and pfdhps

SP has been adopted and implemented in India for treating malaria cases in 1995. The frequent pfdhfr and pfdhps mutants were reported in Delhi in 1994 (e.g. A16C50I51C59N108I164) and Odisha in 1995 (I431A436A437E540A581A613, I431S436G437E540A581A613 and I431A436A437K540A581A613). These findings indicate that these current validated mutations associated with SP resistance were already present before 1995. This result could be likely due to drug pressure exerted by SLP drug which was prescribed in country till 1995 (Fig. 1).

Double mutations in pfdhfr were predominant in this study, and this finding is consistent with that of previous systematic review that double pfdhfr mutations are dominant in India with overall proportion of 57% [25]. Also, we found that R59N108 double mutation was most frequently seen in all pfdhfr double and triple mutants reported in present study. Only R59N108L164 and I51R59N108 triple mutants were reported in India so far [52]. Two new triple mutants (i.e., I51N108L164 and I51R59L164) were found in P. falciparum isolates from Mewat (Haryana).

Regarding pfdhps, double mutations were most frequently seen in contrast to other reports that found a predominance of triple mutations in India [25]. Due to high circulation of SP-resistant P. falciparum populations in NE states (Assam, Mizoram, Meghalaya, Manipur, Nagaland, Tripura, Arunachal Pradesh), AS + SP has been replaced by AL as current malaria treatment [53,54,55]. The pfdhps A436G437E540 triple mutations are highly prevalent in these states, and the findings from the study support this fact as 55.6% of isolates carrying these three mutations were seen in Meghalaya samples. It is noteworthy that another triple mutation (A436G437N540) was found in P. falciparum isolates collected from Assam in the year 1999. The first description of this triple mutant in India was documented in 2005 in isolates from The Nicobar Islands [56]. Thus, the present study confirms that this A436G437N540 triple mutant was circulating in India before 2005.

Finally, this study reports high rates (34.7%) of pfdhfr-pfdhps quadruple mutants but quintuple and sextuple mutants at fewer rates (11.3% and 3.2%). One isolate from Delhi showed quintuple mutations, I51R59N108–G437E540, associated with fully resistance to SP, and strong predictor of clinical SP treatment failure [57, 58]. Imported malaria and SLP/SP drug pressure could likely explain such high rates of drug resistance mutants in pfdhfr/pfdhps. SLP, a sulfamide analogue is used 1982 in India for treatment of CQ-resistant P. falciparum malaria based on satisfactory evidence from clinical studies [59], while SP was implemented 13 years later. Also, most pfdhfr/pfdhps samples (70.3%) analysed were collected in 1994 and 1995, and drug pressure exerted by SLP through cross-resistance mechanism could explain drug resistance mutations found in these samples collected before official implementation of SP. Finally, SP pressure and human migration are also additional determinants of drug resistance in pfdhfr/pfdhps samples by exerting drug pressure and bringing mutant alleles from other countries, respectively.

Pfmdr1

The pfmdr1 SNPs at 86Y and 184F were found in the isolates at varying proportions in study sites consistent with previous studies that reported high proportions of pfmdr1 86Y mutation in West Bengal, Chhattisgarh, and Odisha [42,43,44], but contradicting findings were reported from Mizoram, Meghalaya, and Tripura [60]. In general, the 86Y and 184F mutations are more commonly seen in Asian and African settings, while the 1034C, 1042D and 1246Y mutations are more frequent in South America settings [61]. There is strong link between pfmdr1 86Y mutation and acquisition of resistance phenotype to CQ and AQ, while sensitivity phenotype to DHA, MEF and L [61]. Evidence for involvement of 184F mutation is still limited [61, 62].

Pfk13

No mutations associated with ART-resistance were found in this study. This is in line with previous studies, even though four ART-resistance validated mutations (i.e., 446I, 539T, 561H, 625R) have been reported at very marginal proportions in two areas of India (Arunanchal Pradesh, West Bengal) [63,64,65]. Other pfk13 polymorphisms (189T, 481V, 533A, 549Y, 578S, 579T, 657H, 672S, 675V, and 702N) have been also reported in these two regions and elsewhere (Mizoram, Tripura, Madhya Pradesh, and Assam) [63, 64, 66,67,68]. This implies that ART-resistance has not yet emerged, but molecular surveillance should be continuously carried out in malarious regions in India.

Profile of anti-malarial drug resistance P. vivax genes

Pvcrt-o

Previous in vitro studies reported link between the pvcrt-o K10 insertion and decreased susceptibility of P. vivax parasites to CQ [38], thereby suggesting possible role of this mutation in modulating P. vivax susceptibility. In the present study, all isolates were wild type consistent with previous reports from Thai–Cambodian border, Thailand and China–Myanmar border [69,70,71]. However, this finding is not consistent with that of other studies conducted in India which reported K10 proportion of ~ 9.5–17.5% and 5.6% in Chandigarh (North India) and Mangalore (South India), respectively [72, 73]. Likewise, higher proportions were reported in other endemic regions such as Myanmar (~ 28.2–72.7%) and China–Myanmar border (33.2%) [74,75,76]. All these findings indicate a spatiotemporal variation of K10 proportion in P. vivax areas.

Pvdhfr and pvdhps

Double and triple/single mutants in pvdhps accounted for ~ 33% and ~ 60% of all mutants across India [25]. In this study, double and single mutants were found at overall proportion of 11.9% and 17%, respectively. Double mutants R58N117 were predominant in this study, especially in Delhi, as also previously reported from southern and western parts of India (i.e., Tamil Nadu, Karnataka, and West Bengal) [77,78,79]. This finding is also consistent with other P. vivax endemic countries such as China–Myanmar border, Ethiopia, and Sudan [75, 80, 81]. Imwong and colleagues showed the pvdhfr R58 and N117 mutations are the first to appear when drug pressure is applied [82], and this could likely explain high rates of these mutations in India where SP pressure is high due to utilization of the ACT AS + SP as nationwide first-line treatment of uncomplicated malaria with the exception of NE states. No triple and quadruple mutants were found in this study, which is consistent with previous studies in India that indicates probably a focused geographical distribution of these mutants in NE states of the country [25]. On analysis of the pvdhps gene, the wild type S382A383M399K512R525A553K555V585A661 was predominantly seen in P. vivax isolates, and this is line with the current situation on pvdhps genetic profile in India [25]. Only two types of pvdhps mutants were reported in the present study (SGKAV and SGNAV), and these were also reported in earlier reports from Delhi, West Bengal, Karnataka, Rajasthan, and Tamil Nadu [25]. For the first time this study reports the presence of double mutant SGNAV in Goa samples  from India.

Pvmdr1

The pvmdr1 908L, 958M and 1076L mutations were highly prevalent in P. vivax samples with proportion of 100% in most study areas. Such findings were also reported previously in India, Ethiopia, Pakistan, and China–Myanmar border, but much lower proportions were reported from Thailand [69, 70, 72, 73, 76, 83, 84]. The 976F mutation, found associated with CQ resistance in vitro [38], was not found among P. vivax isolates and this supports earlier findings of low prevalence of this mutation from settings, such as India (7%) and China–Myanmar border (~ 0–2.7%) [70, 72, 73, 75], but contrasting with those from Thailand (~ 1.7–26.7%) and Indonesia (~ 66.7–96.1%) [69, 74, 85].

Pvk12

To the best knowledge of authors, this is first study on pvk12 polymorphism from India. A limited polymorphism was found in pvk12, the P. vivax orthologue gene of pfk13, with one novel nonsynonymous polymorphism (K264R) in isolates from Goa. Other nonsynonymous mutations have been reported from SEA and Oceania areas namely N57I, M124I, S452R, R501K, V541A, E553K, C566G (China–Myanmar border), I537V (Vanuatu), V552I (Cambodia, Malaysia), M548I (Thai–Cambodian border), G581R (China), K596R and P641L (Thai–Cambodian border), and V652L (Solomon Islands) [71, 86,87,88,89,90,91]. In contrast, other studies from SEA (China–Myanmar border) and Africa (Mauritania) reported no pvk12 polymorphism in P. vivax isolates collected from local and imported malaria patients [76, 92, 93]. It is still elusive if these mutations within and outside the pvk12 propeller domain can modulate P. vivax susceptibility to ACT, thereby requiring further research.

Limitations

This study should be interpreted in light of its limitations. First, samples were not from all Indian regions and this limits the representativeness of results at national level. Second, not all samples could be sequenced analysed in this study. Finally, due to low number of good quality sequences for some study sites and year of sample collection, it was impossible to apply sophisticated statistical methods such as generalized equation models to analyse the evolution of resistance mutation over time.

Conclusions

The profiling of genetic markers associated with P. falciparum and P. vivax drug resistance was determined over a 30-year timeframe in India. The analysis revealed substantial spatiotemporal changes with increase in SNPs related to genetic profile of anti-malarial drug molecular markers in P. falciparum and P. vivax populations over 30 years. These findings support continuous surveillance and characterization of P. falciparum and P. vivax populations as proxy of the effectiveness of anti-malarial drugs in India.