|Home | About | Journals | Submit | Contact Us | Français|
TROCAI is a phenotypic tropism test developed using the virological response to a short-term exposure to maraviroc monotherapy (Maraviroc Clinical Test [MCT]). It was found that with TROCAI, a cutoff of <0.5% of dual/mixed viruses was needed to predict R5 HIV tropism. Here, we have validated TROCAI, using this cutoff, in a new cohort of 42 patients, finding a very high concordance between TROCAI and MCT (98%), and a good concordance (71 to 87%) with other genotypic/phenotypic methods.
Determining HIV coreceptor usage is essential before prescribing the CCR5 antagonist maraviroc (MVC) (1, 2). The most widely used coreceptor tropism test is the recombinant enhanced-sensitivity Trofile phenotypic assay (ESTA; Monogram Biosciences) (3). However, this phenotypic assay has some limitations, such as a significant rate of nonreportable results, and specimens must be shipped to the unique reference laboratory in the United States (4, 5). Consequently, other clinical (6), phenotypic (7), and genotypic (8) alternatives for determining viral tropism have been examined. We consider that the virological response to the drug, without the presence of other active drugs with a potential confounding effect that could affect the virological response, should be the most important criteria in order to decide on an MVC prescription. The Maraviroc Clinical Test (MCT) is an in vivo drug sensitivity test based on the virological response to a short-term exposure to MVC monotherapy to be used prior to a recommendation of CCR5 antagonist therapy in both naive and treatment-experienced patients (9, 10). Using the MCT model, we developed a phenotypic tropism test (TROCAI) that overcomes some of the ESTA limitations and obtained a strong correlation between a cutoff of <0.5% dual/mixed viruses and sensitivity to MVC (11).
Here, our aim was to validate TROCAI, using the above-mentioned cutoff, in a new cohort of patients, comparing the TROCAI results with those of the MCT, together with other genotypic/phenotypic methods (see the supplemental material).
Baseline characteristics (just before MVC administration) of the patients are shown in Table S1 in the supplemental material. TROCAI obtained an R5 result (X4 strains, <0.5%) in 35/42 (83%) patients, whereas dual/mixed (DM) results (≥0.5% X4 strains) were obtained in 7/42 (17%) of them (Table 1). Virological response to MVC exposure was observed in 34 (81%) patients (MCT+), while 8 (19%) showed no virological response (MCT−). Interestingly, the number of HIV RNA copies in the cell line U87-CD4+ CXCR4+ (log viral load [VL] U87X4) was statistically different between the MCT− and MCT+ groups (5.5 copies/ml [4.2 to 7.4 copies/ml] versus 2.8 copies/ml [2.3 to 3.3 copies/ml], respectively; P < 0.001) but not in the cell line U87-CD4+ CCR5+ (6.6 copies/ml [5.1 to 8.8 copies/ml] versus 6.3 copies/ml [5.7 to 7.1 copies/ml], respectively; P = 0.848). In the multivariate logistic analysis including the variables associated in the unadjusted analysis, after controlling for potential confounders and avoiding collinearity, log VL of U87X4 was the only variable independently associated with a virological response to MVC (see Table S2 in the supplemental material). We performed a pairwise comparison of concordance between tropism methods (see Table S3 in the supplemental material). The highest concordance was found between TROCAI and MCT (97.6%), followed by the g2p algorithm optimized with the results from MOTIVATE studies with 2% and 5.75% false-positive rate (g2p MOT) and TROCAI (86.8%), and g2p MOT and MCT (84.2%). All TROCAI results but one were concordant with MCT. Patient 15 (P15) did not show sensitivity to MVC exposure, being classified as MCT−; however, the percentage of X4 strains found by TROCAI was 0.01%. In this patient, we obtained in plasma the amino acid sequence 3′-CTRPNNNTRKGIHIGPGRAFYTTGH/RIIGDIRQAHC-5′. We found a mixture of two basic amino acids (histidine [H] and arginine [R]) at position 25 (underlined in the sequence shown above). This amino acid position has been associated with X4 tropism. Interestingly, this sequence was classified as DM by the g2p algorithm recommended by the European Consensus with false-positive rate of 10% (g2p EUR), as R5 by PSSM, and as R5 by g2p MOT but with a warning note recommending additional phenotyping. We also sequenced the virus from cell culture used to perform TROCAI, obtaining the same sequence. We performed an analysis of sensitivity to MVC (see the supplemental material). We obtained a 50% inhibitory concentration (IC50) 7.7-fold higher for the discordant virus strain isolated from P15 than for virus strain BaL (10,318 versus 1,335) (Fig. 1). Immunovirological outcome after the short-term MVC treatment was optimal for all MCT− and MCT+ patients (data not shown).
We have performed the validation of TROCAI by categorizing and comparing the results with the virological response to MVC in a new cohort of patients. Tropism prediction by TROCAI showed a high concordance with MCT, confirming it to be more reliable than other tropism methods for predicting MVC sensitivity. The potential reasons for the lower correlation between the other assays, such as ESTA, and MCT could be because TROCAI was developed using MCT as a reference instead of the results from MVC clinical trials (MOTIVATE/A4001029 and MERIT), where a potential confounding effect of the other active drugs accompanying MVC is possible (12). In addition, TROCAI is not based on the construction of recombinant viruses bearing a representative env gene, or a part thereof, of the patient's HIV population, as other tests do (7). This point is also different from most of the genotypic tropism tests where only the V3 loop is analyzed to predict HIV tropism. The log VL of U87X4 reported by TROCAI was the only variable independently associated with the virological response to MVC after controlling for potential cofounders, confirming previous results (13). A possible increase of X4-tropic strains due to the effect of MVC therapy has been suggested (14); however, no deleterious effects on the immunovirological outcome after the short-term MVC exposure have been observed either in this study or in others (10, 15, 16). In addition, in the only discordance between TROCAI and MCT, we found a particular virus carrying the amino acid mutation 25H/R. It has been communicated that this particular mutation does not discriminate between the X4 and R5 phenotypes (17), so we cannot discard that other mutations outside the V3 loop could be associated (18) with the lower sensitivity to MVC observed in our study.
Our study has some limitations. First, the traditional notion is that extended virus culturing during virus stock production could bias the composition of the quasispecies assessed (2). However, we previously explored this issue, showing that multiple cycles of infection do not change virus tropism, at least during the time period needed to perform TROCAI (11). Second, the success of TROCAI viral stock production decreases to 50% in patients with a VL of <500 HIV RNA copies/ml. Nevertheless, this limitation could be enhanced by introducing improvements to the culture conditions (19, 20). Third, the number of patients included in this validation cohort was relatively low, but was higher and with different patient characteristics than those of the initial cohort used to develop TROCAI.
In summary, we have performed a validation of the phenotypic test TROCAI, categorizing and comparing its tropism prediction with the virological response to MVC. TROCAI is a reliable tool to predict MVC sensitivity, showing a high concordance with other clinical, phenotypic, and genotypic methods.
We are grateful to all the patients who have participated in this study and to Marien Sanchez and Magdalena Rodriguez for their clinical support. Samples from patients were kindly provided by the HIV BioBank integrated in the Spanish AIDS Research Network (RIS).
This work was supported by Redes Telemáticas de Investigación Cooperativa en Salud (RETICS; 2006, Red de SIDA RD06/0006/0021, 2007–2010 and RD16/0025/0020) and Pfizer/ViiV Healthcare (grants WS843473 and WS2425049). A.G.-S. is supported by Instituto de Salud Carlos III (CD14/00320). E.R.-M. has a grant from Fondo de Investigaciones Sanitarias (CD014/0025).
We declare no conflicts of interest.
This work was supported by Redes Telemáticas de Investigación Cooperativa en Salud (RETICS; 2006, Red de SIDA RD06/0006/0021, 2007–2010, and RD16/0025/0020) and Pfizer/ViiV Healthcare (grants WS843473 and WS2425049).
Supplemental material for this article may be found at http://dx.doi.org/10.1128/AAC.01326-16.