Plasmodium falciparum artemisinin-resistant K13 mutations confer a sexual-stage transmission advantage that can be overcome with atovaquone-proguanil

24 Containing the spread of artemisinin (ART)-resistant Plasmodium falciparum will be 25 assisted by improved u nderstanding of its human-to-mosquito transmission. We 26 compared gametocyte dynamics among field isolates containing K13 mutations 27 conferring ART resistance and K13 wild-type parasites. In Pailin, Cambodia, the male 28 to female gametocyte ratio was higher among K13 mutant infections compared to K13 29 wild-type infections. We also investigated the effects of artesunate and atovaquone -30 proguanil on the transmissibility of an ART-resistant K13 mutant strain, Cam3.IIR539T, 31 in a volunteer infection study. Gametocyte production was higher after a single dose 32 of artesunate (2 mg/kg) in volunteers infected with ART-resistant compared to ART-33 sensitive parasites. Despite the presence of gametocytes in volunteers infected with 34 ART-resistant parasites, there was no infection observed in Anopheles stephensi 35 mosquitoes after atovaquone-proguanil treatment. We report transmission 36 determinants of ART-resistant infections that could be advantageous over ART -37 sensitive infections. Moreover, we show additional benefits of treating ART-resistant 38 infections with atovaquone-proguanil treatment. 39 MAIN TEXT 40

41 The appearance and spread of Plasmodium falciparum strains resistant to artemisinin 42 (ART) derivatives, the main component of first-line antimalarial therapies, threatens 43 malaria control and elimination . Infections caused by ART-resistant parasites are 44 characterised by slower parasite clearance after treatment and the presence of point 45 mutations such as C580Y, Y493H, and R539T in the beta-propeller domain of the P. 46 falciparum Kelch13 (K13) protein (PF3D7_1343700)1. ART-resistant infections were 47 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 3 first reported in western Cambodia and the Thailand–Cambodia border in 2002–20042. 48 Over the past 10 to 15 years, ART-resistant parasites have come to dominate across 49 the Greater Mekong Subregion 1. Recently, K13 mutat ions associated with ART 50 resistance have also been detected in other areas such as Guyana, Papua New 51 Guinea and Rwanda3-5. 52 Higher levels of gametocytes - the parasite stage infectious to mosquitoes - have been 53 suggested among patients infected with ART-resistant P. falciparum parasites1. These 54 resistant parasites might, therefore, have higher transmissibility. However, because 55 gametocyte levels can be affected by different variables including genetic 56 background6, accumulation of gametocytes after a long asymptomatic infection7, and 57 antimalarial drug exposure 8-11, a thorough examination of gametocyte dynamics in 58 ART-resistant infections is needed. 59 In addition, antimalarials with transmission -blocking activities against ART-resistant 60 strains are needed to slow down or stop the ir spread. Detailed investigation of the 61 effects of antimalarials on gametocyte dynamics, using data from field studies , has 62 been hampered by several limitations. These include the inability to ascertain the 63 duration of infection s prior to individuals seeking treatment, the difficulty of directly 64 estimating the sexual commitment rate (the number of young gametocytes produced 65 per asexual parasite cycle), and the paucity of data on male to female gametocyte 66 ratios12. 67 Volunteer infection studies are playing an increasingly important role in accelerating 68 antimalarial drug development, and in improving our understanding of P. falciparum 69 gametocyte dynamics13-15. In the Induced Blood Stage Malaria (IBSM) model, healthy 70 volunteers are inoculated intravenously with blood-stage Plasmodium parasites and 71 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 4 monitored closely using molecular assays. In addition, xenodiagnosis by direct skin 72 feeding or membrane feeding assays, using blood samples collected from infected 73 volunteers, can be performed to determine infectiousness to mosquitoes 13,16. 74 Recently, we have reported an IBSM model using the ART-resistant P. falciparum K13 75 mutant Cam3.IIR539T strain17. 76 Herein, w e used data from two complementary studies to test for differences in 77 gametocyte dynamics, including male to female gametocyte ratio s and gametocyte 78 levels, in patients and volunteers infected with K13 mutant and K13 wild -type P. 79 falciparum parasites. We have also leveraged the ART-resistant IBSM model to 80 compare the effect of a single dose of artesunate on gametocyte production in 81 volunteers infected with ART-resistant or ART-sensitive strains and to investigate the 82 effect of atovaquone-proguanil (AP) treatment on the transmissibility of ART-resistant 83 parasites. 84

85 Variation in gametocyte dynamics across field sites from the TRAC study. 86 We used the mean log2 expression ratio of 121 late-stage gametocyte-specific genes 87 to determine gametocyte levels in 771 clinical samples collected from 11 field sites, as 88 this strongly correlated with the ratio of gametocyte to total parasite numbers as 89 counted by microscopy across all sites (r=0.73, p<0.0001) (Supplementary Fig 1A ). 90 Gametocyte transcript levels varied significantly among field sites (Kruskal-Wallis test, 91 p<0.0001) (Fig. 1A). Samples collected from Pailin, Cambodia had statistically 92 significant higher levels of gametocyte transcripts than those originating from Mae Sot 93 in Thailand, Attapeu in Laos, or Ramu in Bangladesh (Fig. 1A), Supplementary Table 94 1). Gametocyte levels were also higher in samples from the Democratic Republic of 95 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 5 the Congo compared to Bangladesh , despite both sites having only K13 wild -type 96 isolates (Fig. 1A-B). These findings suggest that genetic background is an essential 97 contributor to transmission potential. A considerable variation in the male to female 98 gametocyte ratio across and within the different sites was observed, possibly reflecting 99 variation observed during the course of an infection14 (Supplementary Fig. 1 B and 100 Supplementary Table 1). A significant positive but weaker association was also 101 observed between asexual parasite density and the male to female gametocyte ratio 102 (r=0.23, p=0.005; (Supplementary Fig. 1C). When we stratified the gametocytes and 103 male to female gametocyte ratios by sampling site, we found in Pailin that infections 104 with K13 mutants had slightly lower gametocyte gene expression levels than infections 105 with wild-type parasites (~1.3-fold, adjusted p=0.03; Fig. 2 A-B). However, the ratio of 106 male to female gametocytes was significantly higher in K13 mutant than wild -type 107 infections by ~2 -fold (adjusted p=0.004; Fig. 2C, Supplementary Table 1). This 108 suggests that the Pailin K13C580Y parasites may be more infectious to mosquitoes and 109 support higher rates of outbreeding and transmission. 110 Effect of artesunate on the emergence of gametocytes in the K13 versus 3D7 111 study. 112 Artesunate treatment initially reduced the total parasitaemia in all volunteers infected 113 with either ART-resistant (n=13) or ART-sensitive parasites (3D7) (n=9) (Fig. 3A). 114 ART-resistant infected volunteers experienced recrudescence earlier (Day 11) than 115 ART-sensitive infected volunteers (between Day 15 and Day 20). 116 Accordingly, the peak of female and male gametocytemia occurred earlier among 117 volunteers infected with ART-resistant parasites. Female gametocytemia peaked at a 118 median of 12 days post -artesunate (range: 10-14 days) and male gametocytemia at 119 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 6 14 days post-artesunate (range: 10-18 days) in volunteers infected with ART-resistant 120 parasites, while in volunteers infected with ART -sensitive parasites the female and 121 male gametocytaemia both peaked at 17 days post -artesunate (female range:15-21 122 days and male range:14-21 days); p<0.001) (Figure 3B-C and Supplementary Fig. 3). 123 A positive correlation between total parasitaemia pre -artesunate and total 124 gametocytemia 12 days post -artesunate was observed in volunteers infected with 125 ART-resistant parasite s (Supplementary Fig. 2B and Supplementary Table 2). In 126 contrast, no correlation between total parasitaemia AUC pre -artesunate and total 127 gametocytemia AUC 12 days post -artesunate was observed in volunteers infected 128 with ART-sensitive parasites (Supplementary Fig. 2B and Supplementary Table 2). 129 These data suggest, as previously reported, that artesunate inhibits the development 130 of young gametocytes in ART-sensitive strains18. 131 The male to female gametocyte ratio varied over time in each volunteer 132 (Supplementary Figure 4). We log10-transformed the ratio and calculated the mean in 133 three periods: 22 days 134 post-inoculation, and then compared between groups. Overall, the male to female ratio 135 was higher among volunteers infected with ART-resistant than ART -sensitive 136 parasites (0.200 vs 0.096, p=0.004; Supplementary Table 4), translating to ratios of 137 2:10 and 1:10 male to female gametocytes among volunteers with ART-resistant 138 versus ART-sensitive infections respectively (Supplementary Table 4). 139 Effect of artesunate on gametocyte production in volunteers infected with ART-140 resistant parasites. 141 The in vivo sexual commitment rate did not differ significantly between volunteers 142 infected with ART -resistant versus ART -sensitive parasites before artesunate 143 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 7 treatment (Day 9), (median: 0.66, [range: 0.58-0,70; n=13] vs 0.63, [range: 0.55-0.67; 144 n=3]). However, after artesunate treatment (Day 11) volunteers infected with ART -145 resistant parasites experienced recrudescence and had significantly higher sexual 146 commitment rates during recrudescence after artesunate treatment (Day 11) 147 compared to before treatment (Day 9) (Fig. 4A-B). In contrast, ART-sensitive parasite 148 infections did not recrudesce by Day 11 and young gametocytes were not detected 149 after artesunate treatment (Day 11) in volunteers infected with ART -sensitive 150 parasites. 151 After 15 days of ex vivo culture, samples from volunteers infected with ART-resistant 152 parasites taken after artesunate treatment (Day 11) yielded mo re mature male and 153 female gametocytes than samples taken before artesunate treatment (Day 9) (p=0.02, 154 Fig. 5A).This supports our findings that more young gametocytes were present post -155 artesunate treatment than pre-artesunate in the ART-resistant infected volunteers. 156 Moreover, the proportion of male gametocytes was similar in parasite cultures taken 157 pre- and post-artesunate treatment (pre-treatment: 0.9–1.2% vs post-treatment: 0.9–158 1.4%) in volunteers infected with ART-resistant parasites (Fig. 5B), suggesting that 159 artesunate did not affect male to female gametocyte ratios. 160 Before artesunate treatment (Day 9) , blood samples taken from volunteers infected 161 with ART-resistant and ART-sensitive parasites yielded a relatively similar number of 162 gametocytes (ARTr: 2.1%, 2.2% and 3.9%, Fig 5A) vs (ARTs: 0.5%, 2.7% and 4.2%, 163 Supplementary Fig 5A), and with a similar male to female gametocyte ratio (Fig 5B 164 and Supplementary Fig 5B). However, samples taken from volunteers infected with 165 ART-sensitive parasites after artesunate treatment (Day 11) did not develop 166 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 8 gametocytes, consistent with our data showing a loss of viable rings on Day 11 167 following treatment. (Supplementary Fig. 5A). 168 Effect of AP on the infectiousness to Anopheles mosquitoes of volunteers with 169 ART-resistant parasite infection. 170 Transmissibility of volunteers to An. stephensi mosquitoes was analysed by DFA and 171 eMFA in 12/13 volunteers infected with ART -resistant parasites (Supplementary Fig. 172 6). 173 Among volunteers that did not receive AP treatment prior to conducting the 174 transmission assays (n=7), four were infectious to An. stephensi mosquitoes by DFA 175 (57%), and six by eMFA (86%). The median mosquito infection rate among volunteers 176 infected with ART-resistant parasites by DFA was 5% (range: 3–10%, n = 4) and by 177 eMFA was 74% (range: 20–97%; n = 6) (Supplementary Fig. 7A). The four successful 178 infections by DFA had significantly higher male (Mann-Whitney U test; p= 0.03) and 179 female (Mann-Whitney U test; p= 0.03) gametocyte levels than those that were not 180 infectious on the day when transmission assays were performed (Supplementary Fig. 181 7B). 182 Among volunteers that receive d AP treatment prior to conducting the transmission 183 assays (n=5), only two were infectious to mosquitoes, one by DFA with an infection 184 rate of 4% and one by eMFA (Supplementary Fig. 8), with an infection rate of 1 .3%. 185 Interestingly, these five volunteers also had the highest levels of male and female 186 gametocytes among the volunteers infected with ART -resistant parasites at the time 187 of the transmission assays (Supplementary Fig. 8). Thus, in the last cohort of the trial, 188 we decided to evaluate the effect of AP on transmission by conducting a paired 189 comparison, pre- and post -AP treatment with three volunteers infected with ART-190 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 9 resistant parasites. The m osquito infection rates for the three volunteers pre-AP 191 treatment were 0%, 3% and 6.7% (n = 30 mosquitoes) by DFA (Fig. 6A) and 10% (n 192 = 39 mosquitoes), 80% (n=50 mosquitoes) and 91% (n = 47 mosquitoes) by eMFA 193 (Fig. 6B). The infection rates for two volunteers post-AP treatment were all negative 194 by DFA (Fig. 6A), and only one was positive 2% (n= 50 mosquitoes) by eMFA (Fig. 195 6B). No significant difference in gametocytemia was found between pre- and post-AP 196 among ART-resistant infected volunteers (female: pre-AP median: 2,569 [range: 1,115 197 – 11,449] vs post-AP median: 3,449 [range: 475-3,551] gams/mL; p= 0.58; male: pre-198 AP median 978 [range: 236-987] vs post-AP median: 670 [range: 150-1282] gams/mL; 199 p=0.67) (Fig. 6A). 200

201 In this study, we investigated the transmissibility of parasites carrying K13 mutations, 202 with the underlying hypothesis that increased transmissibility may be a factor in the 203 spread of ART-resistant strains across the Great er Mekong Subregion. Using an 204 experimental infection model with Cam3IIR539T parasites carrying the K13 mutant 205 allele, we showed that a single dose of artesunate increased the sexual commitment 206 rate in volunteers infected with these parasites compared to volunteers infected with 207 the ART-sensitive reference strain 3D7. 208 Moreover, a survey of male and female gametocyte transcriptomes obtained from 209 patients from various geographical regions revealed a higher ratio of male to female 210 transcriptional signatures particularly in K13 mutants from Pailin, Cambodia, where 211 artemisinin resistance is well established. Although higher overall gametocyte levels 212 have been shown to contribute to increased transmissibility to Anopheles 213 mosquitoes19, mathematical modelling has also predicted that a high male to female 214 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 10 gametocyte ratio in low -density infections could enhance the chances of successful 215 transmission20,21. Thus, the higher male to female gametocyte ratio among K13 -216 mutant infections compared to K13 wild-type in Pailin, a hot -spot of ART resistance, 217 suggests that this could be a contributor to the spread of resistance in the region. 218 We also assessed, for the first time , the in vivo effect of artesunate on the sexual 219 commitment rate of ART-resistant and ART-sensitive parasites. In this study, a single 220 dose of artesunate cleared ART-sensitive asexual parasitemia to a level where young 221 gametocytes could no longer be detected but enhanced the sexual commitment rate 222 in infections with the ART-resistant parasite. This is a key transmission advantage in 223 areas where ART resistance is not present and therefore transmission -blocking 224 antimalarials, such as primaquine, are not routinely administere d22. Similarly, an 225 increase on gametocytemia, after a 6-hour DHA pulse in vitro assay with 700 nM DHA, 226 has been reported in a P. falciparum strain with a different k13 allele conferring 227 artemisinin resistance (Pf IPC-5202C580Y)11. Although in vitro assays measuring sexual 228 commitment rates have been described, the lengthy protocols make them impractical 229 for large field studies23,24. Thus, quantification of the dynamic changes in the sexual 230 commitment rate of parasites in field studies ha s been scarce. However, a new 231 approach combining ex vivo culture and molecular assays targeting young 232 gametocytes such as pfap2g and pfgdv1 proved useful in characterising sexual 233 commitment rates in isolates from Ghana 25. A field study characterising sexual 234 commitment rates in areas of ART resistance would contribute to our understanding 235 of the overall transmission potential of these strains. 236 Likewise, considering the evidence on sexual commitment plasticity i n a murine P. 237 chabaudi parasite model in response to environmental factors (e.g. drop in 238 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 11 haematocrit, drug pressure) 26 and the results presented here in, further investigation 239 on the effect of current antimalarial regime s on gametocyte production in ART-240 resistant strains is necessary. 241 Although we did not investigate the mechanism by which artesunate enhanced the 242 sexual commitment rate in ART-resistant infections, a number of possible mechanisms 243 would be worth investigat ing in future studies. For instance, assessing whether 244 artesunate treatment results in changes in host-factors such as 245 lysophosphatidylcholine (LysoPC) content in the plasma27, thereby modulating sexual 246 commitment in the surviving parasites. Another possible explanation is that the ART 247 resistance mechanisms of these strains may have a direct or indirect influence on the 248 sexual commitment rate of drug-resistant parasites. ART-resistant parasites undergo 249 metabolic changes resulting in reduced haemoglobin endocytosis28. It is possible that 250 the lack of amino acids could generate a general cellular stress response in the 251 parasite29,30. Thus, gametocytogenesis might be triggered as a collateral effect of the 252 perceived starvation state in the parasite26. 253 This study used the ART -sensitive strain 3D7 . Ideally, the comparator strain for the 254 investigations performed herein should be an isogenic parasite such as the rev erted 255 K13 wild-type Cambodian strain (Cam3.IIrev) with an identical genetic background as 256 the K13 mutant , Cam3.IIR539T used in this study . Unfortunately, logistic al difficulties 257 inherent in producing a clinical-grade bank of such parasites enabling a clinical trial 258 prevented us from including this control experiment. Hence, we cannot exclude the 259 possibility that other genetic differences between the Cam3.IIR539T parasite and the 260

strain 3D7 may contribute to the observed effect in sexual commitment 261 rates. These results may be explained by in vitro data, where other non-K13 262 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 12 polymorphisms modulate sexual commitment rates and thus may explain the observed 263 higher level of gametocytemia in areas with slower parasite clearance after 264 treatment1,6. More research is required to fill the gaps in identifying determinants of 265 transmission of ART -resistant parasites in endemic settings , including the genetic 266

of the parasites. 267 Using a combination of molecular tools and ex vivo culture, we found that artesunate 268 treatment did not substantially affect the proportion of male gametocytes in ART-269 resistant infections. These results support those of Witmer et al .9, whereby ART-270 resistant male gametocytes carrying the k13 R539T and C580Y mutations from 271 Southeast Asia were not affected by artemisinin derivatives. This suggests that under 272 artemisinin drug pressure , a n ART-resistant strain that is a naturally higher male 273 gametocyte producer will continue to transmit better than a sensitive strain, which is a 274 significant concern considering the observed male gametocyte bias seen in ART -275 resistant K13 mutant isolates from Pailin, Cambodia. 276 To prevent the spread of ART-resistant parasites in areas where they are prevalent , 277 the WHO recommends giving a single low-dose of primaquine (SLD PQ, 0.25mg/kg 278 or 15mg ), the only widely available antimalarial that targets mature gametocytes, 279 added to a standard antimalarial treatment31. However, new antimalarials with 280 transmission-blocking effect are needed as SLD PQ can still cause haemoglobin loss 281 in G6PD deficient patients22. We showed that a 3 -day course of AP was sufficient to 282 prevent the infection of Anopheles mosquitoes. T hese results support further 283 assessment of atovaquone or AP as transmission-blocking antimalarial in areas where 284 ART resistance is prevalent. In support of these data, atovaquone was recently shown 285 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 13 to synergize with dihydroartemisinin against Cam3.IIR539T parasites, while not showing 286 this synergy against Cam3.IIWT parasites (Mok et al., manuscript under review). 287 The transmission -blocking effect of AP has been established in vitro with an 288 artemisinin-sensitive strain in standard MFAs 38. Atovaquone was also effective at 289 blocking P. falciparum transmission when applied ectopically to infected Anopheles 290 mosquitoes39. However, this is the first time that the transmission-blocking effect of AP 291 has been tested in a n experimental human infection model with an ART-resistant 292 parasite. Although there has been a reluctance to use this drug combination, due to 293 the rapid development of atovaquone resistance in the parasite32,33, evidence showing 294 that such mutations preclude onward transmission of resistant parasites in the 295 mosquito34 suggests further consideration of this approach. 296 We note that in this study the low number of transmission events observed in the 297 clinical trial did not allow us to determine whether enhanced sexual commitment rates 298 after drug exposure in volunteers infected with ART-resistant parasites translated into 299 increased infecti vity overall. However, s everal studies have reported enhanced 300 infectivity in resistant parasites after treatment 35-38. Other uncontrolled variables 301 including gametocyte maturity39,40, human host genetic factors41-44, and immune 302 responses45,46, could significant ly influence the transmission potential of a P. 303 falciparum infection. 304 In conclusion, this study suggests that ART -resistant infections have a greater 305 transmission potential than ART-sensitive strains. Thus, these data have implications 306 in areas where ART resistance is emerging, as it shows how sub-curative doses of an 307 artemisinin component can favour the propagation of isolates with resistant 308 phenotypes. Moreover, it highlights the importance of considering ART -resistant 309 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 14 parasites from various geographical backgrounds when developing antimalarials that 310 target P. falciparum sexual stages , as their effectiveness can vary based on the 311 parasite’s genetic makeup. Finally, this study suggests an additional benefit of a 3 -312 dose AP treatment as a transmission-blocking antimalarial in areas where ART 313 resistance is prevalent. 314

315 Clinical trials 316 Tracking Resistance to Artemisinin Collaboration I (TRAC I) study. 317 Transcriptome data previously generated for the clinical malaria samples that were 318 collected under the multi -collaborative TRAC I study between 2011 and 2013 were 319 used in this analysis 1,29. The TRAC I microarray data is accessible through GEO 320 Series accession number GSE59099 321 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE59099). At the time of 322 collection, ART resistance was either established (in Pursat and Pailin, Cambodia; 323 Mae Sot and Ranong, Thailand; and Binh Phuoc, Vietnam), newly emerging (in Preah 324 Vihear and Rattanakiri, Cambodia; Attapeu, Laos; and Shwe Kyin, Myanmar) or 325 absent (in Ramu, Bangladesh; and Kinshasa, Democratic Republic of Congo (DR 326 Congo)) 1. Sites with fewer than 15 isolates were excluded (Sisakhet and Khun Han 327 in Thailand; and Ilorin in Nigeria). We also removed isolates that showed 328 heterozygosity or missing genotype data in the k13 allele1. In total, we select ed 771 329 out of 1048 clinical samples from 10 sampling sites in Southeast Asia and one in 330 Africa. 331 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 15 Briefly, whole blood was collected from patients arriving at health facilities upon 332 admission. After white blood cell depletion using CF11 columns, RNA was extr acted 333 using TRIzol® 29. cDNA was synthesized and amplified using a 5’ -3’ reverse -334 transcriptase template switching-PCR protocol (SMART) and then hybridized against 335 a reference pool consisting of 3D7 asexual blood stages on a custom long -336 oligonucleotide microarray chip as describ ed previously 29. The average log 2 337 expression ratios of 121 highly abundant gametocyte stage IV and V genes were used 338 as an index to quantify gametocytes present in the samples 47. The male to female 339 gametocyte ratio was estimated by calculating the ratio of the average transcript 340 abundance of six male-specific gene markers to the average transcript abundance of 341 14 female-specific gene markers (listed in Supplementary Table 5). To ensure that this 342 analyses was based on samples with either female or male gametocytes present, we 343 only considered samples that showed average transcript abundance in either male or 344 female gametocyte genes of at least 1.42-fold higher than the asexual reference pool 345 (log2 expression >0.5) (Supplementary Table 1). This totalled 139 samples. 346 This study was approved by the relevant local ethics committees and the Oxford 347 Tropical Research Ethics Committee. All adult patients or the parents of children gave 348 written informed consent. 349 The ART-resistant IBSM model. 350 An IBSM volunteer infection study was undertaken 17 to compare parasite clearance 351 after single -dose artesunate in volunteers infected with either ART -resistant 352 Cam3.IIR539T parasites or the ART -sensitive reference strain 3D7 17. This clinical trial, 353 referred to herein as t he K13 vs 3D7 trial, was a phase 1, open -label trial with 3 354 cohorts, conducted at Q-Pharm Pty Ltd, Brisbane, Australia17. 355 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 16 Thirteen healthy malaria -naïve volunteers were inoculated with ART -resistant 356 parasites (Cam3.II R539T) on Day 0, a nd nine volunteers with 3D7 ART -sensitive 357 parasites on Day 1. Parasitaemia was monitored by qPCR of the 18S -rRNA gene48 358 from Day 4 post -inoculation. A single dose of artesunate (2 mg/kg; Guilin 359 Pharmaceutical Shanghai Co. Ltd) was given to all infected volunteers on Day 9 to 360 treat asexual parasitaemia. Piperaquine (960 mg, PCI Pharma Services) was given to 361 all ART -resistant infected volunteers on Day 11 to treat the first recrudescence, 362 whereas for ART -sensitive infected volunteers piperaquine was given between Day 363 15 and Day 20 when asexual parasitemia recrudesced (Supplementary Table 3). AP 364 (Malarone®, GlaxoSmithKline Australia) was given between Day 16 and Day 28 to 365 treat the second recrudescence or as a compulsory treatment at the end of the study 366 for all volunteers (Supplementary Table 3). Primaquine was given between Day 21 367 and Day 35 to eliminate gametocytes at the end of the study (Supplementary Table 368 3). 369 Mature female and male gametocyte levels were measured from Day 9 by quantitative 370 reverse transcriptase PCR (qRT -PCR) of transcripts pfs25 and pfMGET (for female 371 and male gametocytes, respectively, and were converted to gametocytes/mL as 372 previously described49. 373 The K13 versus 3D7 trial was approved by the QIMR Berghofer and Australian Red 374 Cross Blood Service Human Research Ethics Committees. All volunteers gave written 375 informed consent before enrolment. The trial was registered with the Australian New 376 Zealand Clinical Trials Registry: ACTRN12617001394336. 377 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 17 Gametocyte dynamics after artesunate treatment in the K13 vs 3D7 trial 378 The area under the curve (AUC) of total parasitaemia and total gametocytemia (sum 379 of female and male gametocytes) was calculated to determine the magnitude of the 380 correlation between them14. The AUC of the total parasite density was calculated using 381 consecutive parasitaemia/mL data from Day 4 until the day of artesunate treatment 382 (Day 9), and the AUC of the total gametocytemia14,49 was from the first positive sample 383 (male and female) until 12 days post artesunate administration (Day 21). This analysis 384 assumed 12 days as the period it would take for gametocytes to mature and return to 385 circulation. 386 The male to female gametocyte ratio was determined by dividing the number of male 387 gametocytes over female gametocytes/mL. For accurate quantification, we only 388 included time points when both male and female gameto cytes were greater than 6 389 gametocytes/mL, which is the lower limit of the reportable range for both gametocyte 390 assays49. 391 Effect of artesunate on sexual commitment rate 392 We evaluated the effect of artesunate on the proportion of young gametocytes 393 produced per asexual cycl e, i.e. the sexual commitment rate. This analysis used a 394 qRT-PCR assay targeting the pfAP2 transcription factor ( pfap2-g, PF3D7_1466400) 395 that is expressed in young gametocyte stages and is required for 396 gametocytogenesis25, as well as the ring-stage transcript skeleton-binding protein 1 397 (pfsbp1, PfE0065w) expressed by asexual ring -stages and young gametocytes 14,50. 398 The sexual commitment rate was defined as the ratio between the log 10-transformed 399 levels of pfap2 and pfsbp1 transcripts, as measured in samples collected from Day 4 400 until Day 12. These time points included the window when young gametocytes stages 401 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 18 induced during artesunate treatment first appeared in the blood circulation (i.e. 402 approximately 48 hours after drug administration)51. 403 An ex vivo approach was used to determine whether artesunate induces a change in 404 the male to female gametocyte ratio. Briefly, b lood samples were taken pre - (Day 9) 405 and post- artesunate treatment (Day 11) from volunteers infected with ART -resistant 406 and ART-sensitive parasites. Asexual parasites were removed with 50 mM of N-Acetyl 407 Glucosamine, (NAG, Sigma-Aldrich), and young gametocytes present in the samples 408 were allowed to mature to stage V under standard in vitro culture conditions (5% 409 haematocrit, RPMI complete media (Sigma -Aldrich), 10% human serum, and a fixed 410 gas composition (5% O 2, 5% CO 2, 90% N 2)52. After 15 days, total parasitemia was 411 measured in the culture pellets by qRT -PCR specific to the 18S -rRNA gene, and 412 gametocyte density was measured by qRT -PCR targeting pfs25 and pfMGET 413 transcripts49 (Fig. 7A). 414 Effect of AP on the transmissibility of ART -resistant infected volunteers to 415 Anopheles mosquitoes. 416 Transmissibility of gametocytes from volunteers infected with ART-resistant parasites 417 was assessed by direct skin feeding assays (DFA) and membrane feeding assays 418 using gametocytes enriched from large blood volumes via Percoll gradient 419 centrifugation (eMFA)14. DFA and eMFA we re performed between Day 21 and Day 420 24, on consenting volunteers presenting with more than 100 female gametocytes/mL 421 on the day of these assays. 422 For DFAs, volunteers were exposed to approximately 30 starved female Anopheles 423 stephensi mosquitoes (Sind-Kasur Nijmegen strain) that were allowed to feed for 15 424 minutes on the forearm. For eMFA, parasite-infected erythrocytes were enriched from 425 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 19 36 mL of whole blood using a 65% Percoll gradient. The gametocyte -enriched layer 426 was mixed with uninfected red blood cells and AB serum (from a malaria-naïve donor, 427 Australian Red Cross Blood Service) and fed to 60 –100 starved mosquitoes for 30 428 minutes using a membrane -feeding apparatus. Blood -fed mosquitoes were returned 429 to controlled environmental conditions (30°C and 70-80% relative humidity with a 12-430 hour day/night cycle) and maintained on a sucrose-PABA solution for eight days14. P. 431 falciparum infection in the mosquito midguts was determined by qPCR targeting the 432 18S-rRNA gene 53. The mosquito infection rate was defined as the proportion of 433 mosquitoes with infected midguts. 434 The effect of AP on tr ansmission was investigated in three volunteers by comparing 435 the gametocyte levels and mosquito infection rates by DFA and eMFA, before and 436 after AP treatment (Fig. 7B). 437 Statistical Methods 438 In the TRAC study, differences in gametocyte gene expression levels between field 439 sites were assessed using Kruskal -Wallis and corrected for multiple comparisons 440 using Dunn’s test. Differences between male to female ratios between the K13 mutant 441 and wild-type samples for each field site was determined by t-test with correction for 442 multiple comparisons using the Holm-Sidak method. 443 For the K13 vs 3D7 study, GraphPad Prism (version 7), was used for linear regression 444 and to compute AUC measures. The pre -treatment parasitaemia AUC and 445 corresponding total gametocytemia AUC were log 10-transformed to assess their 446 correlation. Mann -Whitney U tests were used to i) compare the AUC of total 447 gametocytemia, from first positive until last positive, between ART-resistant and ART-448 sensitive infected volunteers; and ii) compare the day when male or female peak 449 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 20 gametocytemia was reached between ART -resistant and ART -sensitive infected 450 volunteers. R version 3.5.2 was used to calculate the male to female gametocyte ratio. 451 Log10-transformed data and one -way ANOVA tests were used to examine the 452 relationship between male to female gametocyte ratios and day groups. t-tests were 453 used to compare male to female gametocyte ratios between ART-resistant and ART-454 sensitive infections in volunteers. The Wilcoxon’s matched-pairs signed-rank test was 455 used to compare the sexual commitment rate before and after artesunate treatment, 456 as well as gametocyte levels before and after AP treatment. The proportion of female 457 and male gametocytes in culture w as compared using a ratio paired t test. Finally, 458 female and male gametocyte levels between infectious and non-infectious volunteers 459 were compared using the Mann-Whitney U test. 460 DATA AVAILABILITY 461 The authors declare that all other data supporting the findings of this study are 462 available within the article and its supplementary Information files or are available from 463 the authors upon request. 464 465 466 467 468 469 470 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 21

471 The authors thank all volunteers who participated in the study, all staff from the clinical 472 study site Q-Pharm Pty. Ltd and all staff from the QPID Laboratory for PCR analysis. 473 From the QIMR Berghofer Medical Research Institute we thank Matt Adams and 474 Melanie Rampton for technical assistance with mosquito rearing, Katharine Trenholme 475 for assistance with inoculum preparation, Peter O’Rourke for statistical advice, and 476 Laura Cascales for assistance with manuscript writing and editing support. From the 477 University of Melbourne, we thank Pengxing Cao for critical review of the manuscript. 478 From Uniformed Services University of the Health Sciences, Surendra K. Prajapati 479 and Kim Williamson for sharing the protocol to determine sexual commitment rate. 480 This study was supported by funds from Medicines for Malaria Venture 481 (OPP1111147). JSM was supported by an Australian Government National Health and 482 Medical Research Council (NHMRC) Practitioner Fellowship (APP1135955). JSM, MR 483 and ZP were supported by an NHMRC Program G rant (1132975). DAF gratefully 484 acknowledges funding from the NIH (R01 AI109023) and the US Department of 485 Defence (to DAF and SM, W81 -XWH-19-10-086). SM is the recipient of a long -term 486 fellowship from the Human Frontiers of Science Program (LT000976/2016-L). We are 487 extremely grateful to the TRAC clinical site coordinators, doctors, nurses, laboratory 488 technicians, and patients who participated in the multi-site TRAC clinical study. TRAC 489 was funded by the UK Department for International Development (DFID) for the benefit 490 of developing countries and was coordinated by MORU Tropical Health Network which 491 receives core funding from the Wellcome Trust. Whole -genome sequencing and 492 genotyping for TRAC was funded by the Wellcome Trust through core funding of the 493 Wellcome Trust Sanger Institute (098051). 494 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 22 AUTHOR CONTRIBUTION 495 S.M. and D.A.F. conducted the transcriptomic analysis; K.A.C., R.E.W. J.J.M. and 496 J.S.M. designed the K13 vs 3D7 clinical trial; G.J.R. and C.Y.T.W. performed 497 molecular analysis in the K13 vs 3D7 clini cal trial.; Z.P., H.M., S.L. and G.J.R. 498 performed mosquitoes assays; J.G., M.R. and Z.P. conducted culturing experiments; 499 Z.P., L.W., S.M. and L.M. performed formal statistical analysis; A.O., B.B. and J.S.M. 500 conducted clinical assessment of volunteers; and J.S.M., M.W.A.D. and Z.P. designed 501 the transmission experiments. Z.P., S.M., J.S.M. and D.A.F. wrote the manuscript, and 502 all authors read the manuscript and approved its submission. 503 COMPETING INTERESTS 504 I have read the journal’s policy and the authors of this manuscript have the following 505 competing interests: Z.P., K.A.C., M.R., R.E.W., G.J.R., H.M., S.L., J.G., L.W., Sam 506 M., A.O., B.B., L.M., and J.S.M. are employees of the study sponsor QIMR Berghofer 507 Medical Research Institute. J.J.M. is an employee of the Medicines for Malaria 508 Venture, which provided funding for the study. 509

& CORRESPONDENCE 510 Correspondence to Zuleima Pava at [email protected]; James McCarthy at 511 [email protected]; David Fidock at [email protected] and 512 Sachel Mok at [email protected]. 513 514 515 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 23

516 1 Ashley, E. A. et al. Spread of artemisinin resistance in Plasmodium falciparum 517 malaria. N Engl J Med 371, 411-423, doi:10.1056/NEJMoa1314981 (2014). 518 2 Dondorp, A. M. et al. Artemisinin resistance in Plasmodium falciparum malaria. 519 N Engl J Med 361, 455-467, doi:10.1056/NEJMoa0808859 (2009). 520 3 Mathieu, L. C. et al. Local emergence in Amazonia of Plasmodium falciparum 521 k13 C580Y mutants associated with in vitro artemisinin resistance. Elife 9, 522 doi:10.7554/eLife.51015 (2020). 523 4 Miotto, O. et al. Emergence of artemisinin -resistant Plasmodium falciparum 524 with kelch13 C580Y mutations on the island of New Guinea. bioRxiv, 621813 525 (2019). 526 5 Uwimana, A. et al. Emergence and clonal expansion of in vitro artemisinin-527 resistant Plasmodium falciparum kelch13 R561H mutant parasites in Rwanda. 528 Nat Med, doi:10.1038/s41591-020-1005-2 (2020). 529 6 Eksi, S. et al. Plasmodium falciparum gametocyte development 1 (Pfgdv1) and 530 gametocytogenesis early gene identification and commitment to sexual 531 development. PLoS Pathog 8, e1002964, doi:10.1371/journal.ppat.1002964 532 (2012). 533 7 Barry, A. et al. Increased gametocyte production and mosquito infectivity in 534 chronic versus incident Plasmodium falciparum infections. medRxiv (2020). 535 8 Lozano, S. et al. Gametocytes from K13 propeller mutant Plasmodium 536 falciparum clinical isolates demonstrate reduced susceptibility to 537 dihydroartemisinin in the male gamete exflagellation inhibition assay. 538 Antimicrob Agents Chemother 62, doi:10.1128/AAC.01426-18 (2018). 539 9 Witmer, K. et al. Artemisinin-resistant malaria parasites show enhanced 540 transmission to mos quitoes under drug pressure. bioRxiv, 541 2020.2002.2004.933572, doi:10.1101/2020.02.04.933572 (2020). 542 10 Buckling, A., Ranford -Cartwright, L. C., Miles, A. & Read, A. F. Chloroquine 543 increases Plasmodium falciparum gametocytogenesis in vitro. Parasitology 118 544 (Pt 4), 339-346, doi:10.1017/s0031182099003960 (1999). 545 11 Rajapandi, T. Upregulation of gametocytogenesis in anti-malarial drug-resistant 546 Plasmodium falciparum. J Parasit Dis 43, 458-463, doi:10.1007/s12639-019-547 01110-w (2019). 548 12 Koepfli, C. & Yan, G. Plasmodium gametocytes in field studies: do we measure 549 commitment to transmission or detectability? Trends Parasitol 34, 378 -387, 550 doi:10.1016/j.pt.2018.02.009 (2018). 551 13 Alkema, M. et al. A randomized clinical trial to compare P. falciparum 552 gametocytaemia and infectivity following blood-stage or mosquito bite induced 553 controlled malaria infection. J Infect Dis 222, 1416 -1416, 554 doi:10.1093/infdis/jiaa157 (2020). 555 14 Collins, K. A. et al. A controlled human malaria infection model enabling 556 evaluation of transmission-blocking interventions. J Clin Invest 128, 1551-1562, 557 doi:10.1172/JCI98012 (2018). 558 15 Reuling, I. J. et al. A randomized feasibility trial comparing four antimalarial drug 559 regimens to induce Plasmodium falciparum gametocytemia in the controlled 560 human malaria infection model. Elife 7, doi:10.7554/eLife.31549 (2018). 561 16 Beshir, K. B. et al. Residual Plasmodium falciparum parasitemia in Kenyan 562 children after artemisinin -combination therapy is associated with increased 563 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 24 transmission to mosquitoes and parasite recurrence. J Infect Dis 208, 2017-564 2024, doi:10.1093/infdis/jit431 (2013). 565 17 Watts, R. E. et al. Safety and parasite clearance of artemisinin resistant 566 Plasmodium falciparum infection: A pilot and a randomised volunteer infection 567 study in Australia. PLoS Med 17, 568 doi:https://doi.org/10.1371/journal.pmed.1003203 (2020). 569 18 Price, R. N. et al. Effects of artemisinin derivatives on malaria transmissibility. 570 Lancet 347, 1654-1658, doi:10.1016/s0140-6736(96)91488-9 (1996). 571 19 Churcher, T. S. et al. Predicting mosquito infection from Plasmodium falciparum 572 gametocyte density and estimating the reservoir of infection. Elife 2, e00626, 573 doi:10.7554/eLife.00626 (2013). 574 20 Bradley, J. et al. Predicting the likelihood and intensity of mosquito infection 575 from sex specific Plasmodium falciparum gametocyte density. Elife 7, 576 doi:10.7554/eLife.34463 (2018). 577 21 Paul, R. E., Brey, P. T. & Robert, V. Plasmodium sex determination and 578 transmission to mosquitoes. Trends Parasitol 18, 32 -38, doi:10.1016/s1471 -579 4922(01)02122-5 (2002). 580 22 Bancone, G. et al. Single low dose primaquine (0.25 mg/kg) does not cause 581 clinically significant haemolysis in G6PD deficient subjects. PloS One 11, 582 e0151898, doi:10.1371/journal.pone.0151898 (2016). 583 23 Brancucci, N. M., Goldowitz, I., Buchholz, K., Werling, K. & Marti, M. An assay 584 to probe Plasmodium falciparum growth, transmission stage formation and 585 early gametocyte development. Nat Protoc 10, 1131 -1142, 586 doi:10.1038/nprot.2015.072 (2015). 587 24 Lucantoni, L., Duffy, S., Adjalley, S. H., Fid ock, D. A. & Avery, V. M. 588 Identification of MMV malaria box inhibitors of Plasmodium falciparum early-589 stage gametocytes using a luciferase-based high-throughput assay. Antimicrob 590 Agents Chemother 57, 6050-6062, doi:10.1128/AAC.00870-13 (2013). 591 25 Usui, M. et al. Publisher Correction: Plasmodium falciparum sexual 592 differentiation in malaria patients is associated with host factors and GDV1 -593 dependent genes. Nature Commun 10, 2740, doi:10.1038/s41467-019-10805-594 w (2019). 595 26 Schneider, P. et al. Adaptive plasticity in the gametocyte conversion rate of 596 malaria parasites. PLoS Pathog 14, e1007371, 597 doi:10.1371/journal.ppat.1007371 (2018). 598 27 Brancucci, N. M. B. et al. Lysophosphatidylcholine regulates sexual stage 599 differentiation in the human malaria parasite Plasmodium falciparum. Cell 171, 600 1532-1544 e1515, doi:10.1016/j.cell.2017.10.020 (2017). 601 28 Birnbaum, J. et al. A Kelch13 -defined endocytosis pathway mediates 602 artemisinin resistance in malaria parasites. Science 367, 51 -59, 603 doi:10.1126/science.aax4735 (2020). 604 29 Mok, S. et al. Drug resistance. Population transcriptomics of human malaria 605 parasites reveals the mechanism of artemisinin resistance. Science 347, 431-606 435, doi:10.1126/science.1260403 (2015). 607 30 Tilley, L., Straimer, J., Gnadig, N. F., R alph, S. A. & Fidock, D. A. Artemisinin 608 action and resistance in Plasmodium falciparum. Trends Parasitol 32, 682-696, 609 doi:10.1016/j.pt.2016.05.010 (2016). 610 31 GMP, W. Updated WHO policy recommendation: single dose primaquine as a 611 gametocytocide in Plasmodium falciparum malaria. Geneva: WHO (2012). 612 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 25 32 Cottrell, G. et al. Emergence of resistance to atovaquone -proguanil in malaria 613 parasites: insights from computati onal modeling and clinical case reports. 614 Antimicrob Agents Chemother 58, 4504 -4514, doi:10.1128/AAC.02550 -13 615 (2014). 616 33 Staines, H. M. et al. Clinical implications of Plasmodium resistance to 617 atovaquone. Antimicrob Agents Chemother 73, 581 -595, 618 doi:10.1093/jac/dkx431 (2018). 619 34 Goodman, C. D. et al. Parasites resistant to the antimalarial atovaquone fail to 620 transmit by mosquitoes. Science 352, 349-353, doi:10.1126/science.aad9279 621 (2016). 622 35 Hogh, B. et al. The differing impact of chloroquine and 623 pyrimethamine/sulfadoxine upon the infectivity of malaria species to the 624 mosquito vector. Am J Trop Med Hyg 58, 176 -182, 625 doi:10.4269/ajtmh.1998.58.176 (1998). 626 36 Ecker, A., Lakshmanan, V., Sinnis, P., Coppens, I. & Fidock, D. A. Evidence 627 that mutant PfCRT facilitates the transmission to mosquitoes of chloroquine -628 treated Plasmodium gametocytes. J Infect Dis 203, 228 -236, 629 doi:10.1093/infdis/jiq036 (2011). 630 37 Mharakurwa, S. et al. Malaria antifolate resistance with contrasting Plasmodium 631 falciparum dihydrofolate reductase (DHFR) polymorphisms in humans and 632 Anopheles mosquitoes. Proc Natl Acad Sci U S A 108, 18796 -18801, 633 doi:10.1073/pnas.1116162108 (2011). 634 38 Lambrechts, L., Halbert, J., Durand, P., Gouagna, L. C. & Koella, J. C. Host 635 genotype by parasite genotype interactions underlying the resistance of 636 anopheline mosquitoes to Plasmodium falciparum . Malar J 4, 3, 637 doi:10.1186/1475-2875-4-3 (2005). 638 39 Hallett, R. L. et al. Chloroquine/sulphadoxine-pyrimethamine for gambian 639 children with malaria: transmission to mosquitoes of multidrug -resistant 640 Plasmodium falciparum . PLoS Clin Trials 1, e15, 641 doi:10.1371/journal.pctr.0010015 (2006). 642 40 Targett, G. et al. Artesunate reduces but does not prevent posttreatment 643 transmission of Plasmodium falciparum to Anopheles gambiae . J Infect Dis 644 183, 1254-1259, doi:10.1086/319689 (2001). 645 41 Gouagna, L. C. et al. Genetic variation in human HBB is associated with 646 Plasmodium falcipa rum transmission. Nat Genet 42, 328 -331, 647 doi:10.1038/ng.554 (2010). 648 42 Joice, R. et al. Inferring developmental stage composition from gene expression 649 in human malaria. PLoS Comput Biol 9, e1003392, 650 doi:10.1371/journal.pcbi.1003392 (2013). 651 43 Graves, P. M., Carter, R., Burkot, T. R., Quakyi, I. A. & Kumar, N. Antibodies to 652 Plasmodium falciparum gamete surface antigens in Papua New Guinea sera. 653 Parasite Immunol 10, 209 -218, doi:10.1111/j.1365 -3024.1988.tb00215.x 654 (1988). 655 44 Graves, P. M., Doubrovsky, A., Satt abongkot, J. & Battistutta, D. Human 656 antibody responses to epitopes on the Plasmodium falciparum gametocyte 657 antigen PFS 48/45 and their relationship to infectivity of gametocyte carriers. 658 Am J Trop Med Hyg 46, 711-719, doi:10.4269/ajtmh.1992.46.711 (1992). 659 45 Usui, M. et al. Plasmodium falciparum sexual differentiation in malaria patients 660 is associated with host factors and GDV1 -dependent genes. Nature Commun 661 10, 2140, doi:10.1038/s41467-019-10172-6 (2019). 662 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 26 46 Bousema, T. et al. Human immune responses that reduce the transmission of 663 Plasmodium falciparum in African populations. Int J Parasitol 41, 293 -300, 664 doi:10.1016/j.ijpara.2010.09.008 (2011). 665 47 Young, J. A. et al. The Plasmodium falciparum sexual development 666 transcriptome: a microarray analysis using ontology -based pattern 667 identification. Mol Biochem Parasitol 143, 67 -79, 668 doi:10.1016/j.molbiopara.2005.05.007 (2005). 669 48 Rockett, R. J. et al. A real -time, quantitative PCR method using hydrolysis 670 probes for the monitorin g of Plasmodium falciparum load in experimentally 671 infected human volunteers. Malar J 10, 48, doi:10.1186/1475 -2875-10-48 672 (2011). 673 49 Wang, C. Y. T. et al. Assays for quantification of male and female gametocytes 674 in human blood by qRT -PCR in the absence of p ure sex-specific gametocyte 675 standards. Malar J 19, 218, doi:10.1186/s12936-020-03291-9 (2020). 676 50 Tadesse, F. G. et al. Molecular markers for sensitive detection of Plasmodium 677 falciparum asexual stage parasites and their application in a malaria clinical 678 trial. Am J Trop Med Hyg 97, 188-198, doi:10.4269/ajtmh.16-0893 (2017). 679 51 Bancells, C. et al. Revisiting the initial steps of sexual development in the 680 malaria parasite Plasmodium falcip arum. Nat Microbiol 4, 144 -154, 681 doi:10.1038/s41564-018-0291-7 (2019). 682 52 Farid, R., Dixon, M. W., Tilley, L. & McCarthy, J. S. Initiation of 683 gametocytogenesis at very low parasite density in Plasmodium falciparum 684 infection. J Infect Dis 215, 1167-1174, doi:10.1093/infdis/jix035 (2017). 685 53 Wang, C. Y. T. et al. Assessing Plasmodium falciparum transmission in 686 mosquito-feeding assays using quantitative PCR. Malar J 17, 249, 687 doi:10.1186/s12936-018-2382-6 (2018). 688 689 690 691 692 693 694 695 696 697 698 699 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 27 FIGURE LEGENDS 700 Figure 1. Levels of gametocyte transcripts in symptomatic clinical samples 701 collected from the TRAC study in 2011 -2013. Levels of gametocyte transcripts in 702 symptomatic clinical samples collected from the TRAC study in 2011-2013. a) Box and 703 whisker plots represent the median and 10 -90 percentile of the mean of 121 704 gametocyte gene log 2 expression ratios in 771 clinical isolates across 11 field sites. 705 *** adjusted p<0.001, **** adjusted p<0.0001. Only statistically significant pairwise 706 comparisons with adjusted p<0.001 as calculated by Dunn’s multiple comparisons test 707 following Kruskal-Wallis test are shown. b) Proportion of K13 mutant genotypes among 708 the 771 clinical samples. 709 Figure 2. Distribution of gametocyte levels and male to female gametocyte ratios 710 by K13 genotype across 11 field sites . a) Number of K13 wild -type or mutant 711 samples in each field site (N=771). b) Box and whiskers represent the median and 10-712 90 percentile of gametocyte transcript levels stratified by K13 genotype status in each 713 field site for 771 clinical isolates. This analysis suggested a slightly lower gametocyte 714 prevalence for K13 mutants relative to wild -type isolates in Pailin, Cambodia. c) Box 715 and whiskers represent the median and 10-90 percentile of male to female gametocyte 716 ratios stratified by K13 genotype status (red – mutant; black - wild-type), in each site 717 for 139 clinical samples ( where the average male or female gametocyte transcript 718 levels were higher than the asexual reference pool by ≥1.42-fold). Numbers listed are 719 the number of clinical samples for K13 mutant or wild -type used in the analyses. Of 720 the five field sites analyzed, the data showed a 2 -fold higher proportion of male 721 gametocyte in K13 mutants relative to wild -type isolates in Pailin, Cambodia. * 722 adjusted p<0.05, ** adjusted p<0.01. See Supplementary Table 2 for transcript levels 723 in isolates. As a point of reference, the ratio of male to female gametocytes for in vitro 724 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 28 cultured 3D7 mature gametocytes profiled by the same microarray platform is 0.69 725 which is lower than the clinical samples. 726 Figure 3. Parasitaemia counts before and after artesunate and piperaquine 727 treatment. a) Total parasitemia, b) female, and c) male gametocytem ia counts in 728 volunteers infected with ART -resistant (left) and ART -sensitive (right) parasites. 729 Purple and green lines indicate the time points when artesunate and piperaquine were 730 administered, respectively. AP and primaquine treatment days are described in 731 Supplementary Table 3. Thick coloured lines are the median values per day and thin 732 grey lines are the individual values; non-detect data points were substituted by 1. 733 Figure 4. Sexual commitment rate increased after artesunate treatment in ART-734 resistant infected volunteers. a) Sexual commitment rate (ratio of log10 gametocyte 735 ring transcript pfap2g / log10 ring transcripts pfsbp1) over time, as measured by qRT-736 PCR. The purple dashed line indicates administration of 2 mg/kg of artesunate; the 737 green dashed line indicates administration of 960 mg/kg of piperaquine; the solid black 738 line is the median of pfap2g/pfsbp1 ratio per day; and grey lines are the pfap2g/pfsbp1 739 ratio for each participant. b) Sexual commitment rate was significantly higher on Day 740 11 than on Day 9 or 12 (Wilcoxon -paired test). Colours indicate each of the 13 ART -741 resistant infected volunteers. AS: artesunate; PIP: piperaquine. 742 Figure 5. Ex vivo gametocyte cultures from post-artesunate samples developed 743 more gametocytes than th ose from pre -artesunate samples, in volunteers 744 infected with ART-resistant parasites. a) Proportion of total gametocytes (male and 745 female) developed after 15 days of ex vivo culture as measured by qRT-PCR for pfs25 746 and pfMGET, and normalized by total parasitaemia (18S rRNA), in samples taken pre 747 and post artesunate treatment (ratio paired t-test); b) Bar -plots showing the sex 748 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 29 composition of the ex vivo cultures 15 days after sample collection on Day 9 and Day 749 11. 750 Figure 6. Prevalence of mosquito infection pre and post -atovaquone-proguanil 751 (AP) treatment on Day 21 in ART -resistant (ARTr) volunteers. a) Infections were 752 carried out by Direct Feeding Assay (DFA) and green bars indicate the proportion of 753 infected midguts, red circles indicate female gametocytemia (gametocytes/mL), and 754 blue circles indicate male gametocytemia (gametocytes/mL) detected by qRT-PCR at 755 the time of feeding, n= number of mosquitoes used in each assay b) Infections were 756 carried out by enriched Membrane Feeding Assay (eMFA) and blue bars indicate the 757 proportion of infected midguts, n= number of mosquitoes used per assays AP. NP: 758 Not performed. 759 Figure 7. Study designs using the ART -resistant IBSM model a) ex vivo 760 gametocyte production in ART -resistant (ARTr) and ART-sensitive (ARTs) infected 761 volunteers. b) effect of atovaquone -proguanil (AP) on transmission in three ART -762 resistant volunteers enrolled in the K13 versus 3D7 clinical trial. 763 All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 0 1 2 Mean of gametocyte gene log2 expression ratio *** **** **** **** ******* *** 0 50 100 No. of samples wild-type Y493H R539T I543T C580Y A675V other mutations Pu rsat, Cambo dia Preah V ihear ,Camb odia Rattan akir i, C amb od ia Pai lin,Camb odia Mae So t, T hailan d Rano ng, T hailan d Binh Phu oc, Vietnam Bag o D ivision, M ya nma r Attap eu,Laos Ramu, Banglad esh Kinshasa, DR C on go Field site Figure 1. a. b. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint -1 0 1 2 3 Mean of gametocyte gene log2 expression ratio wild-type mutant Pu rsat, Cambo dia Preah V ihear ,Camb odia Rattanakir i, C amb od ia Pai lin,Camb odia Mae So t, T hailan d Ran ong, T hailan d Binh P hu oc, Vi etn am Bag o D ivision, M ya nma r Attap eu,Laos Ramu, Banglad esh Kinshasa, D R C on go 0.0 0.5 1.0 1.5 2.0 2.5 Field site Ratio of male to female gametocyte gene expression 0 20 40 60 80 100 No. of samples wild-type mutant * ** 4 14 4 5 12 0 14 0 3 23 0 2 10 6 15 5 5 0 6 0 11 0 a. b. c. Figure 2. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 100 101 102 103 104 105 106 107 Parasites/mL 100 101 102 103 104 105 106 Female gametocytes/mL - 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 100 101 102 103 104 105 106 Study Day Male gametocytes/mL - ART-resistant infected volunteers Artesunate treatment day Piperaquine treatment day 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 Study Day ART-sensitive infected volunteers a. b. c. Figure 3. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint Pre-AS (Day 9) Post-AS (Day 11) Post-AS (Day 12) 0.0 0.2 0.4 0.6 0.8 1.0pfap2g / pfsbp1 p=0.01 p<0.001 p= 0.001 2 3 4 5 6 7 8 9 10 11 12 13 0.0 0.2 0.4 0.6 0.8 1.0 Study Day pfap2g / pfsbp1 AS PIP b.a. Figure 4. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 0 20 40 60 Proportion of total gametocytes produced ex vivo (%) Pre-Artesunate Post-Artesunate (Day 9 cultures) (Day 11 cultures) p = 0.02 ARTr-11ARTr-12ARTr-13 ARTr-11ARTr-12ARTr-13 0 20 40 60 80 100 Proportion of male/female gametocytes produced ex vivo (%) FemaleMale Pre-Artesunate Post-Artesunate (Day 9 cultures) (Day 11 cultures) a. b. Figure 5. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint 0 20 40 60 80 100 100 101 102 103 104 105 Prevalence of infection (% mosquitoes infected) Gametocytes/mL 0% NP0% 0% 3.3% 6.7% Pre Post Pre Post Pre Post n=30 n=30 n=30 n=30 n= 30 n=30 ARTr-11 ARTr-12 ARTr-13 ARTr-11 ARTr-12 ARTr-13 0 20 40 60 80 100 Prevalence of infection (% mosquitoes infected) 0% NP2% 10% 91% 80% Pre Post Pre Post Pre Post n=39 n=50 n=47 n=50 n= 50 n=50 Direct Feeding Assay Enriched Membrane Feeding Assay a. b. Figure 6. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint a. b. Figure 7. All rights reserved. No reuse allowed without permission. perpetuity. preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in The copyright holder for thisthis version posted October 27, 2020. ; https://doi.org/10.1101/2020.10.26.20214619doi: medRxiv preprint