Systematic Review and Meta-Analysis on Human African Trypanocide Resistance

Background Human African trypanocide resistance (HATr) is a challenge for the eradication of Human African Trypansomiaisis (HAT) following the widespread emergence of increased monotherapy drug treatment failures against Trypanosoma brucei gambiense and T. b. rhodesiense that are associated with changes in pathogen receptors. Methods: Electronic searches of 12 databases and 3 Google search websites for human African trypanocide resistance were performed using a keyword search criterion applied to both laboratory and clinical studies. Fifty-one publications were identified and included in this study using the PRISMA checklist. Data were analyzed using RevMan and random effect sizes were computed for the statistics at the 95% confidence interval. Results: Pentamidine/melarsoprol/nifurtimox cross-resistance is associated with loss of the T. brucei adenosine transporter 1/purine 2 gene (TbAT1/P2), aquaglyceroporins (TbAQP) 2 and 3, followed by the high affinity pentamidine melarsoprol transporter (HAPT) 1. In addition, the loss of the amino acid transporter (AAT) 6 is associated with eflornithine resistance. Nifurtimox/eflornithine combination therapy resistance is associated with AAT6 and nitroreductase loss, and high resistance and parasite regrowth is responsible for treatment relapse. In clinical studies, the TbAT1 proportion of total random effects was 68% (95% CI: 38.0–91.6); I2 = 96.99% (95% CI: 94.6–98.3). Treatment failure rates were highest with melarsoprol followed by eflornithine at 41.49% (95% CI: 24.94–59.09) and 6.56% (3.06–11.25) respectively. HATr-resistant phenotypes used in most laboratory experiments demonstrated significantly higher pentamidine resistance than other trypanocides. Conclusion: The emergence of drug resistance across the spectrum of trypanocidal agents that are used to treat HAT is a major threat to the global WHO target to eliminate HAT by 2030. T. brucei strains were largely resistant to diamidines and the use of high trypanocide concentrations in clinical studies have proved fatal in humans. Studies to develop novel chemotherapeutical agents and identify alternative protein targets could help to reduce the emergence and spread of HATr.


Introduction
The World Health Organization (WHO) has set 2030 as a target for the elimination of human African trypanosomiasis (HAT) [1]; however, the development of drug-resistant phenotypes (see [2] on trypanocide resistance) in resource-poor countries affected by HAT presents a major challenge for HAT control. In Africa, HAT is caused by Trypanosoma brucei gambiense (TBG, chronic variant) and T. b. rhodesiense (TBR, acute variant), also referred to as gHAT and rHAT, respectively. Uganda has the misfortune to harbor both sub-species within its borders [3]. Trypanosomes are able to evade host immune defenses through a process of antigenic variation (while each genome contains over 10 3 distinct variable surface glycoprotein (VSG) genes, every trypanosome typically expresses a single VSG that rapidly switches [4], compromising any efforts to develop vaccines).
Globally, six major drugs are available for treatment of HAT depending on the stage of the infection: pentamidine, suramin, melarsoprol, eflornithine, nifurtimox/eflornithine combination therapy (NECT), and fexinidazole (see [2] on HAT pharmaceutics and limitations of current approved therapies). The emergence of human African trypanocide resistance (HATr) has undermined the use of monotherapy for HAT treatment.
Pentamidine was the first antiprotozoal diamidine to be routinely used for HAT treatment, its mode of action serving to disrupt the AATT-rich portions of Trypanosoma DNA and suppress mitochondrial activity [5]. Pentamidine resistance has been linked to changes in the transmembrane transport of the drug i.e., Trypanosoma brucei adenosine transporter 1/purine 2 (TbAT1/P2) and high-affinity pentamidine transporter 1 (HAPT1) [5][6][7]. Melarsoprol is an arsenical with which resistance has been associated with aquaglyceroporin transporters 2/3 (AQP2/3) in trypanosomes [8]. Resistance has led to the adoption of combination nifurtimox/eflornithine combination therapy (NECT), which has low toxicity and a shortened therapeutic period, although weekly intravenous infusions have proved challenging in resource-limited settings [9]. In early stages, pentamidine (TBG) and suramin (TBR) are used while melarsoprol (TBG and TBR) and eflornithine (TBG) are recommended for late-stage HAT [3]. New drugs, including pafuramidine maleate for early HAT, have failed, while acoziborole for stage 1 and 2 HAT is under continuous review [10], demonstrating the need to develop novel therapies; for example, fexinidazole is easier to administer since it is an oral medication [11,12]. The objective of the current study was to identify major parasitic markers in HATr associated with clinical studies (field surveys and clinical trials) and experimental (laboratory-based) studies.

Study Design
Multiple electronic databases were searched using the Ovid interface: AMED ( These searches generated a total of 172 publications and the Web of Science generated 111 publications using the following keywords as shown in Supplementary File S1. ALL = ((african trypanosomiasis)) OR ALL = (trypanosoma brucei)) OR ALL = ('tsetse flyborne diseases' OR 'HAT' or 'human African trypanosomiasis'))) AND ALL = ((trypanosoma brucei gambiense) or (trypanosoma brucei rhodesiense))) AND ALL = ((suramin OR melarsoprol OR eflornithine OR nifurtimox OR pentamidine OR (NECT or (Nifurtimox Eflornithine Combination Therapy))))) AND ALL = (trypanocides resistance or drug resistance).
Grey literature was searched on Google using the WHO and CDC websites, and citations in systematic reviews were also searched for papers that might have been missed ( Figure 1).

Article Screening on Inclusion and Exclusion
Data files were exported to EndNote 2020 and all papers were merged (N = 283). The SR depublicator removed 79 duplicates. A total of 204 papers were then exported to Covidence, which removed 3 additional duplicates, and these were confirmed by authors. Title and abstract screening removed 121 papers and only 80 papers were subjected to full-text review. After removing review articles, a total of 46 papers were acquired from the search and an additional 5 papers were added (i.e., 1 from Google search and 4 from review searches), as shown in Figure 1.

Statistical Analysis
Data were exported into MS Excel and categorized into laboratory studies (cell culture and rodents) and clinical studies (involving humans). These data were descriptively presented in tables while quantitative data were analyzed using RevMan ® for meta-analysis using proportions and random effect sizes at 95% confidence interval. Data on the impact of resistance were analyzed using GraphPad Prism version 6 and posthoc Tukey's tests were conducted and represented with different superscripts (a,b) being used to indicate significant differences (p < 0.05)

Description of Human African Trypanocide Resistance in the Study
Pentamidine/melarsoprol cross-resistance (PMXR) is associated with loss of and mutations in TbAT1/P2 genes [13], followed by mitochondrial, post translational activator XAC1, and flagellar genes [14]. The loss of TbAT1 alone does not effectively prevent endocytosis of pentamidine and melarsoprol (see [15], melarsoprol has other entry targets); however, a loss of aquaglyceroprotein (TbAQP, see [16], AQPs are important for viability and osmoregulation) has been associated with complete loss of trypanocide endocytosis [17]. In particular, the loss of TbAQP2/3 following point mutations in TbAT1 leads to a loss of the high-affinity pentamidine/melarsoprol transporter (HAPMT) [18,19]. In pentamidine diamidine resistance, HAPT1 and low-affinity pentamidine transporter 1 (LAPT1) are responsible for the residual uptake of melaninophenyl arsenic [20]. The presence of multiple resistance genes in single parasites contributes to a markedly resistant phenotype. For example, TbAT1 and T. brucei multidrug-resistant pentamidine-associated gene (TbMRPA) further complicate melarsoprol resistance [21] (Table 1).
These findings are important since pentamidine and melarsoprol have been the major trypanocides involved in HAT chemotherapy for over 60 years [22]. Cross-resistance between melarsoprol-pentamidine, diminazene aceturate, isometamidium chloride has been reported [23]. Although drugs used in HAT are different from those used for animal African trypanosomiasis, pentamidines/diamidines with similar pathogen targets (TbAT1/P2) contribute to this selection pressure [24]. This is important since TBR and TBG have been isolated from livestock species that have been recognized as maintenance hosts (see [25] in pigs, [26] in small ruminants, and [27] in cattle) and sources of re-infection and introduction of resistant phenotypes in humans [28].
Nifurtimox is metabolized rapidly and its metabolites are not effective against T. brucei s.l., thus reducing its therapeutical effect [29]. Similar results have been reported for TbAT1, where the loss of P2 favors trypanosome survival [30]. The loss of amino acid transporter (AAT) 6 leads to increased eflornithine resistance [24,25] while NECT resistance is associated with multiple loss of function of AAT6 and nitroreductase (NTR) [26,27]. Resistance to arsenicals and diamidine is associated with HAPT loss [31] while TbAT1 loss disrupts the uptake of diamidines [32]. In addition, cross-resistance in arsenical and suramin [33], arsenical/melarsoprol/pentamidine and diminazene aceturate [15] as well as isometamidium and diminazene aceturate, which are mainly used in livestock [34], raises major public health concerns since pentamidines and diminazene aceturate are used in both humans and animals [24]. Furthermore, the presence of putative nascent polypeptide associated complex (NAC) isoforms [35] presents a dilemma with opportunities for drug targets and challenges to address HATr (Table 1). Nifurtimox resistance has also been associated with TbAT1/P2 (contrary to previous assumptions [36] in which no relationship was made); however, AAT6 loss has been associated with eflornithine resistance. In NECT resistance, multiple loss of AAT6 and NTR is the hallmark of drug resistance. Nifurtimoxfexinidazole resistance is associated with the rapid metabolism of the sulfoxide and sulfone forms of this compound [22]. Since nifurtimox was first developed for use against American trypanosomiasis [37], it is apparent that further research exploring combination therapy could yield more efficient trypanocides.
Resistance to suramin (developed in 1916) has been associated with switching of VSG expression [31,38]. Furthermore, other mechanisms associated with ATP production, metabolism, cell cycle, and genome segregation [39] will continue to offer opportunities for continued research (Table 1). Although the cellular pathways involved in suramin resistance remain to be discovered (see [40] where resistance was postulated to develop after changes in the drug target by expression of drug extrusion mechanisms), since each parasite contains over 1000 VSG genes, expression switching to one particular VSGSur implies that it is close to impossible to eliminate suramin resistance in a population [41].   Blood-brain barrier via P-glycoprotein and multiple drug resistance-associated protein transporters.
Key: Superscript (*) = strain genus not defined in the article. NA = not applicable.

Human African Trypanocide Resistance in Clinical Studies
In clinical studies, relapse/cure rates have been used to study HATr (see [46] on using DL-α-difluoromethylornithine (DFMO, also referred to as eflornithine in most studies) with suramin against arsenical refractory HAT in mice). This is important since DFMO alone has been associated with resistance [47]. Combination of metronidazole and suramin were used to address arsenical resistance in HATr in Zambia [62], while high doses of nifurtimox have also been explored in the Democratic Republic of the Congo (DRC) [63,64]. Melarsoprol resistance in DRC has been associated with mutations in the AQP2/3 gene [48] and parasite regrowth [65]. Human samples from Tanzania and Ivory Coast showed crossresistance to melarsoprol and suramin due to alterations in TbAT1 [33] or other transport mechanisms [15] (Table 2).
Melarsoprol resistance in Uganda and Tanzania has also been associated with mutations in the TbAT1 gene [66], and high minimum inhibition concentration (MIC) titers (see [49]) in Uganda). In DRC, melarsoprol alone is no longer used in late HAT [67] and eflornithine is now used at this stage [68]. The increasing levels of HATr has led to the promotion of combination therapy [69], as well as capital to invest in the discovery of more potent drugs better than the current drugs on the market.

Human African Trypanocide Treatment Relapse Rates
Treatment failure rates were highest with melarsoprol, followed by eflornithine, 41.49% (95% CI: 24.94-59.09) and 6.56% (3.06-11.25), respectively; however, the reliability of these findings may be biased due to the high I 2 value associated with melarsoprol studies (Table 4). Furthermore, a high level of consistency was associated with nifurtimox studies (I 2 = 0.0%) since this showed a low publication bias.  Treatment relapse rates have been used as indicators of resistance and combination therapies using DFMO-suramin against arsenical/melarsoprol resistance have been explored since eflornithine alone is ineffective [74]. This has continued as an alternative combination for use in late HAT cases because of melarsoprol resistance [75]. To overcome arsenical resistance, combination therapies of metronidazole and suramin have also been used since these are associated with mild side effects when compared with suramin monotherapy (in Zambia), while high doses of nifurtimox have also been used in DRC. Melarsoprol resistance in DRC has been associated with mutations in the AQP2/3 gene [35] and parasite regrowth [56]. Human samples from Tanzania and Ivory Coast have shown cross resistance to melarsoprol and suramin due to alterations in TbAT1 [30] or other transport mechanisms [15].
Melarsoprol resistance in Uganda and Tanzania has also been associated with mutations in the TbAT1 gene [57]. In DRC, melarsoprol alone is no longer used in late-stage HAT [58] and has been substituted with eflornithine for this stage [59]. The increasing levels of HATr have led to the promotion of combination therapy [60], as well as increased capital to invest in the discovery of more potent drugs superior those currently on the market. Furthermore, melarsoprol toxicity and decreasing efficacy has led to phasing out the drug as a frontline treatment against T. b. gambiense; this is now possible with the emergence of potent, safe combination chemotherapies, such as NECT, showing that eflornithine will continue to be around for decades to come. The TbAT1 genotype was the most prevalent marker, although few studies have been conducted in humans on the African continent exploring the genome diversity of HATr. This is important since melarsoprol has the highest treatment failure rates [66,73].

Drug Sensitivity Profiles on HATr Using Resistance Profiling
Some experimental laboratory studies have proved to be unreliable due to low reproducibility, especially when conducting clinical (field-based) studies [76]. This has subsequently given rise to speculative interpretations that make it hard to inform policy [77]. Here, we investigated the level of resistance induced in experimental studies for HATr. The resistance factor was calculated by dividing the IC50 of the resistant population by the IC50 of the wild-type in these studies [13,18,20,29,40]. The mean ± SEM for resistance factors of pentamidine, melarsoprol, suramin, and DB 75 were 84.3 ± 20.12, 12.5 ± 3.0, 8.2 ± 2.0, and 10.8 ± 3.5, respectively ( Figure 2). Furthermore, most laboratory studies have continued to produce strains that are significantly resistant to pentamidines, demonstrating a shift in research direction for the next novel trypanocides. the IC50 of the wild-type in these studies [13,18,20,29,40]. The mean ± SEM for resistance factors of pentamidine, melarsoprol, suramin, and DB 75 were 84.3 ± 20.12, 12.5 ± 3.0, 8.2 ± 2.0, and 10.8 ± 3.5, respectively ( Figure 2). Furthermore, most laboratory studies have continued to produce strains that are significantly resistant to pentamidines, demonstrating a shift in research direction for the next novel trypanocides.

Conclusion
Cross-resistance across trypanocides is a major threat to the development of novel monotherapy due to the targeting of similar molecules in the pathogen. TbAT1/P2 are the leading pathogenic transporter targets; however, total resistance is associated with the loss of TbAQ2/3, HAPT1, and LAPT1 in melarsoprol-pentamidine resistance, while AAT6 and NTRs are common in nifurtimox-eflornithine resistance. High treatment failure rates in humans have led to the implementation of high doses, which have proved fatal to patients, highlighting the desperate situation created by HATr in endemic communities.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, File S1: keyword search criteria on HATr.
Author Contributions: K.I.K. and S.C.W. conceptualized the study; K.I.K., S.C.W. and E.T.M. con- Figure 2. Resistance profiles of laboratory-developed HATr phenotypes. Generally, much emphasis has been placed on development of strains heavily expressing resistance to pentamidine compared with all other trypanocides. Different subperscripts in figure 2 are introduced under statistical analysis (i.e., different superscripts (a,b) signify significant differences).

Conclusions
Cross-resistance across trypanocides is a major threat to the development of novel monotherapy due to the targeting of similar molecules in the pathogen. TbAT1/P2 are the leading pathogenic transporter targets; however, total resistance is associated with the loss of TbAQ2/3, HAPT1, and LAPT1 in melarsoprol-pentamidine resistance, while AAT6 and NTRs are common in nifurtimox-eflornithine resistance. High treatment failure rates in humans have led to the implementation of high doses, which have proved fatal to patients, highlighting the desperate situation created by HATr in endemic communities.
Author Contributions: K.I.K. and S.C.W. conceptualized the study; K.I.K., S.C.W. and E.T.M. conducted study design; K.I.K. collected the data; K.I.K. conducted data analysis. All authors have read and agreed to the published version of the manuscript.