Influx of diverse, drug resistant and transmissible Plasmodium falciparum into a malaria-free setting in Gulf Cooperation (GCC) countries. CURRENT STATUS:

Successful malaria control programs have Cooperation Council (GCC) countries. However, a massive influx of imported malaria via migrant workers from endemic areas sustains a threat for the re-introduction of local transmission. Here we examined the origin of imported malaria into one of the GCC countries (Qatar) and assessed the extent of genetic diversity, and carriage of drug resistance genes of imported Plasmodium falciparum and it’s potential to re-introduce the disease. Methods We examined imported malaria reported in Qatar, between 2013 and 2016. We focused on P. falciparum infections and estimated total parasite and gametocyte density using qPCR and qRT-PCR, respectively. In addition, we examined ten neutral microsatellites and four drug resistance genes, Pfmrp1, Pfcrt, Pfmdr1 and Pfkelch13 , to assess the extent of diversity of imported P. falciparum and its potential carriage of drug resistance genotypes respectively. would allow of measures to limit their impact in areas targeted for elimination. The present study examined the source of imported malaria to the transmission-free state of Qatar, and assessed the genetic complexity, drug resistance genes and ability of P. falciparum to produce gametocytes and thus transmit to the vector mosquitoes. Such knowledge would allow control programs to develop targeted policies to reduce circulating parasites, define the source of outbreaks and limit the risk of reintroduction of malaria.

sizes and quantifying peak heights for samples containing multiple alleles per locus. Multiple alleles per locus were scored if electrophoretic peaks corresponding to minor alleles were > 32% the height of the predominant allele [19].

Detection and quantification of early and late P. falciparum gametocytes
Quantitative RT-PCR (qRT-PCR) was used to detect and quantify mRNA from the early gametocytespecific gene, Pfpeg4 [20], and the late gametocyte-specific gene, pfs25 [21].
Total RNA was first isolated from 100 uL of blood taken from a fingerpick, using the SV Total RNA Isolation System (Promega, UK). Quantitative reverse transcription and subsequent amplification (qRT-PCR) of cDNA was carried out using the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher, UK). The RT-PCR conditions and primers used were previously described [17,22].

Amplicon sequencing for characterisation of P. falciparum drug resistance loci
SNPs in four P. falciparum genes, Pfmrp1 (PF3D7_0112200), Pfcrt (PF3D7_0709000), Pfmdr1 (PF3D7_0523000), and PFK13 (PF3D7_1343700), implicated in resistance to several different antimalarial drugs, were typed by amplicon sequence as previously described [23]. The analysis combined multiplex PCR, and custom-designed sequence analysis using Miseq sequencing for high throughput SNP-profiling of drug resistance genes [23]. Seventy P. falciparum isolates were examined. The PfK13 and Pfcrt, genes were each amplified as one fragment, while the longer Pfmdr1 and Pfmrp1 genes were amplified as two fragments. The PCR was carried out in a volume of 25 µl, containing 1 µl (10 pmol) of primers, 0.4 µl of dNTPs (200µlmol/L), 4 µl of Phusion HF buffer (5x) and 1 U of Phusion high-fidelity polymerase enzyme. The cycling temperature profile for all loci was: 98°C/30s, followed by 30 cycles of (98°C/10s, 64°C/4 min), and a final extension at 64°C/5 min The PCR amplicons of all genes for each isolate were pooled and purified using Agencourt AMPure XP purification system and quantified using a Qubit double-stranded DNA (dsDNA) HS assay kit (Thermo Fisher Scientific). The sequencing was run on Illumina® MiSeq systems. Initially, libraries were prepared using Nextera XT kit according to the manufacturer's protocol (Illumina Nextera® XT DNA Sample Preparation Guide, 2012). Following PCR cleanup, the libraries were quantified using the Qubit dsDNA BR kit, and evaluated for fragment size using Agilent High Sensitivity DNA Kit for the 2100 Bioanalyzer Instrument (Agilent Technologies, Santa Clara, CA, USA). Each library was normalized for sequencing to 10 pM according to the manufacturer's protocol and following Illumina's (Illumina®, San Diego, CA, USA) technical note for cluster optimization (Illumina Nextera® Library Validation and Cluster Density Optimization, 2013). Sequencing reactions were carried out using the MiSeq Reagent Kit V2 for 50 cycles (MiSeq, Illumina). SNPs in all genes were called using the reference sequence of the P. falciparum 3D7 clone, version 3 (PlasmoDB, PF3D7 v3).

Data Analysis
All samples that contained gametocyte transcripts, as detectable by qRT-PCR, were then analysed for association between gametocyte carriage and total parasitemia. A Mann Whitney U test was used to examine differences between the density of mature and early gametocytes. Spearman's rank order correlation test was used to examine any association between total parasite density (18S rRNA copy number) and the density of either late gametocytes (Pfs25 copy number) or early gametocytes (Pfpeg4 copy number).
Microsatellite allele data were filtered to retain only minor alleles having a peak height of >33 % of the corresponding predominant alleles if more than one allele was present at any locus. Genetic diversity parameters were calculated for the entire dataset using GenAlex v6.5 [24], including polymorphism at each locus as well as overall and within sub-population diversity, between parasites originated from Africa and The Indian subcontinent. Expected heterozygosity was calculated using the formula for 'unbiased heterozygosity' also termed haploid genetic diversity, H e = [n/(n-1)] [1-∑p 2 ] where n is the number of isolates and p the frequency of each different allele at a locus [25]. Multiplicity of infection, defined as the presence of multiple genotypes per infection, was determined by the detection of more than one allele at a locus. To avoid over-estimation of low abundance alleles, only minor alleles having a peak height of >33 % of the corresponding predominant alleles were accepted. The proportion of samples with more than one allele across the ten loci was used to represent the multiplicity of infection (MOI). The maximum number of alleles across the ten loci was used as an index for minimum number of clones per infection (MNC). The overall mean for the index value for each sample was then calculated.

Parasitaemia and gametocytaemia among imported malaria cases
Ninety of the 118 P. falciparum infections were further examined for total parasite and gametocyte density, using qPCR and qRT-PCR, respectively, and diversity of 10 microsatellites and alleles of four genes linked to drug resistance.
The total P. falciparum density among imported cases ranged widely between 32 and 9,218,498 parasites/ml blood with a median 82783 parasites/ml. The median parasite density among imported cases from the Indian Subcontinent (99,572 parasites/ml) was not significantly higher than that from Africa (88,504 parasites/ml) (p=0.394).

Microsatellite polymorphism
All the examined microsatellites were highly polymorphic among P. falciparum isolates originating from both Africa and the Indian Subcontinent ( Table 2). The number of alleles per locus was higher among the African isolates, ranging from 5 for pfg377 to 18 for polyα, compared to the Indian Subcontinent isolates, ranging from 3 for 2490 to 7 for both TA1 and PfPK2 (Table 2; Supplementary   Table 1). However, allelic diversity, summarized as mean expected heterozygosity (He) across the 10 microsatellite loci, was not significantly different among parasites in the Indian Subcontinent (mean He = 0.78) compared to that in Africa (mean He = 0.76).
Multi-locus haplotypes were constructed using predominant alleles at all of the examined loci. All 90 isolates differ from each other in at least one of the examined loci, with exception of two isolates from Africa, sharing an identical haplotype, both from Sudan. Thus, almost every isolate in each of the examined sites carried a unique genotype.

Multiplicity of infection (MOI)
Seventy-six (84.4%) out of the 90 imported P. falciparum isolates with complete set of data harbored multiple genotypes. A similar mean of multiple genotype infections was seen among parasites in the Indian Subcontinent (84.6%) and Africa (84.4%). The minimum number of genotypes per infected person (the mean maximum number of alleles observed at all loci) was slightly lower in Africa (2.16 genotypes) that in the Indian Subcontinent (2.38 genotypes), but this was not statistically significant (P > 0.05).

Genetic differentiation
Alleles of most microsatellites were distributed widely across P. falciparum among imported malaria cases from Africa (n =77) and the Indian Subcontinent (n = 13). However, a large number of private alleles (detected only in one region) were seen in Africa (n = 50) compared to the Indian Subcontinent (n = 5), which may reflect the smaller sample size. Nonetheless, no evidence of genetic differentiation was observed between imported P. falciparum from Africa and the Indian Subcontinent, (F ST =0.055). The genetic relatedness between the P. falciparum populations was further illustrated by PCoA analysis (Figure 2). Analysis of molecular variance (AMOVA) revealed that the majority of the differences were due to variation between individuals within the same group (95%), and only 5% could be attributed to differences between populations.

Distribution of drug resistance genes among imported cases
Seventy imported P. falciparum isolates were examined using amplicon sequencing for four putative drug resistance genes, PfK13, Pfmdr1, Pfcrt and Pfmrp1 (Table 3). With exception of K13, there was no differences in the prevalence of wild type of the examined genes among parasites originated from Africa or Asia. There was a significantly higher prevalence of mutant PfK13 haplotypes among parasites from Africa than Asia. One nonsynonymous mutation in PfK13 (K189T) was observed at a high prevalence (36%) among parasites originating from Sudan, similar to another African countries [28]. Moreover, ten additional nonsynonymous SNPs K108E, L119L, H136N, T149S, K189T, K189N, N217H, R255K, I354V, E433D, G453A, all existed at very low prevalence ranging from 1-3% (Table 3;   Supplementary Table 2).
The dissemination of malaria to transmission-free areas is driven by travelers from endemic areas [35], where a large proportion of the semi-immune inhabitants sustain asymptomatic low levels parasitaemia [36,37]. However, asymptomatic infection can develop to clinical malaria, often after a long period of settlement in a malaria-free country, that can be delayed by as long as 5 years [38]  The above is evident by the high prevalence of gametocyte carriage among imported P. falciparum malaria in Qatar. Fifty-four (74%) out of 73 imported P. falciparum isolates, successfully examined by qRT-PCR, carried gametocyte stages, with a large proportion (37%) harbouring both early and mature gametocytes, indicative of ongoing gametocytogenesis from the asexual population present in the patient, comprising a potential transmission reservoir. qRT-PCR assay of Pfpeg4 and Pfs25 are credible tools for detection of early and late gametocyte stages, respectively [42]. The density of late gametocytes among imported cases, was low compared to that reported in endemic sites, for instance among infected children in Kenya [43]. Nonetheless, low-density gametocytes can readily infect Anopheles, even at sub-microscopical levels [44,45]. Secondary transmission, originating from imported malaria, is often reported in some GCC countries in receptive areas where the Anopheles vector is present, and a favourable ecological habitat prevails [3,46]. The surge in mosquito abundance, in highly seasonal transmission settings, has been linked to upsurge in gametocyte numbers in asymptomatic carriers [21], in line with the hypothesis of enhanced malaria parasite infectivity in response to increased exposure to uninfected mosquitoes at the start of the transmission season [21,47]. Although the chance of resumption of endemic transmission in GCC countries is limited, as a result of effect vector control, the high rate of imported malaria can readily seed outbreaks in receptive areas, if vector control eases [3].
The high level of diversity among imported P. falciparum to Qatar; from Africa (He = 0.76) and the Indian subcontinent (He = 0.78) and genotype multiplicity parallel that reported in local parasites in both sites [48,49], as well as sites where local transmission occurs in the region, in southwest of Saudi Arabia and Yemen [50]. The introduction of novel malaria parasites lineages into the region, via migrants from endemic areas, can enhance the parasite diversity and effective population size (Ne), as there is a direct relationship between the expected level of diversity and Ne [51]. In addition, the combination of high genotype multiplicity and gametocyte carriage, as seen in the present study, increases the likelihood that imported malaria infections will generate novel genotypes, should transmission occur. Assuming that all genotypes are readily transmissible to mosquito, good agreement has been seen between the rate of cross-mating in mosquito and the extent of multiple genotype infections in humans [52]. Thus, the imported malaria represents not only a risk for initiating local transmission, but also the ability to disseminate novel strains that can escape the effect of current drug regimen. Consistent with the above, high prevalence of SNPs in 6 unlinked genes implicated in drug resistance were seen among imported malaria in Qatar. The success of the current efforts to prevent reintroduction of malaria in transmission-free areas, in GCC, relies on effective case management using artemisinin-combination therapy. There is significantly, high prevalence of wild type alleles among parasites in the Indian subcontinent compared to Africa. Nonetheless, known PfK13 mutations C580Y, Y493H, and R539T, associated with slow artemisinin clearance [28] were not detected in imported P. falciparum for both regions. Nonetheless, many low frequency SNPs were seen, while one (K189T) was observed at a high prevalence (36%) among parasites originating from Sudan, which has been reported in other African countries [28]. Nonetheless, parasites carrying this mutation were found to have a similar response to that of wild type parasites [28] and may not impact current ACT regimen in Qatar artesunate/doxycycline or quinine/clindamycin for uncomplicated and complicated/severe P. falciparum malaria, respectively [11]. Therefore, the presence of SNPs linked to tolerance/resistance to artemisinin derivatives and common partners can impede this strategy, and results in persistence and increased parasite reservoir and vulnerability. For example, the high prevalence of the wild-type Pfmdr1 alleles N86 (61%) and D1246 (97%) linked to artemether-lumefantrine (AL) resistance and the N 86 F 184 D 1246 haplotype (43%) associated with AL treatment failure [53] among imported cases, could compromise the efficacy of the first line regimen (AL) for malaria in Qatar. Similarly, the commonly found Pfmrp1 variants F1390 (79%) and I876V (46%), have been linked to decreased susceptibility to artemisinin mefloquine, and lumefantrine [30,54]. Together, these findings demonstrate the possibility of the emergence of ACTresistant parasites that can persist to be transmitted, even following treatment of patients with imported malaria in the region. Previous surveys in Saudi Arabia and Yemen revealed a high prevalence of drug resistance genotypes among locally acquired P. falciparum infection and linked the source of some of them to Africa and Indian Subcontinent [55-57].

Conclusions
The present study highlights the threat of imported malaria for the re-introduction of the disease in receptive malaria-free areas, such as GCC countries. The high genetic diversity and capacity of the imported P. falciparum to produce gametocytes, emphasis the threat of dissemination of drug resistance genotypes, that can escape current control strategies, should local transmission start.
Thus, there is an urgent need for molecular tools for surveillance of imported cases, to limit the risk of re-introduction of malaria in the region.    Figure 1 Correlation between total parasite density with both early gametocyte and late gametocyte density. (A) log total parasitemia (X axis) and log early gametocyte density (Y axis), the fit line in scatter plot shows a weak/non-significant correlation coefficient (r =0.031, p = 0.835). (B) log total parasitemia (X axis) and log late gametocyte density (Y axis), the fit line in scatter plot shows a weak/non-significant correlation coefficient (r= 0.008, p = 0.946).

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.