M. ZINGG¹, A. HAAS¹, N. LABHARDT², T. KLIMKAIT<sup>1

1</sup>Molecular Virology, Department Biomedicine – Petersplatz, University of Basel, Switzerland, ²Swiss Tropical Public Health Institute, Basel

Background:
Knowledge of the host cell tropism of HIV is critical for initiating a therapy with a coreceptor antagonist and may be of importance for judging the clinical course. Genotypic methods are today available that are mainly based on the variable loop 3 (V3) in the viral env gene, either on the plain sequence (Geno2Pheno) or by analyzing duplex properties (XTrack). Since the V3 loop is flanked by sequences with a higher degree of conservation, amplification typically utilizes these regions for PCR amplification. Thereby commonly used primes may reach slightly into the region, and this may potentially introduce a bias, as the primer-provided nucleotides cannot be taken into account for interpretation.

Methods:
To overcome this bias we now designed shorter primers that, through 2’-modifications, are able to retain comparable binding affinity despite the shortening. These primers now do not reach into the V3 loop and allow full representation of patient-derived V3 sequences. The amplified sequences were in parallel analyzed by duplex-tracking (XTrack) and using the genotypic Geno2Pheno (G2P) assay, which is based on tropism prediction using the amino acid sequence of the V3 loop. In silico G2P analysis of V3 sequences differing only in the affected primer region was also performed with consensus sequences from various subtypes.
Replicative phenotyping were used as gold standard in order to clarify contradictory results.

Results:
A comparison of sets of sequences from subtype AE and –C viruses revealed that for most of them there is no principal difference (CXCR4- vs. CCR5-tropism), and no significant differences in the False Positive Rates (FPR) between the two V3 sequence lengths for XTrack or the G2P assays. Surprisingly even the V3-sequence of consensus HIV-2 does not yield a principally different interpretation. Complete omission of the respective flanking sequences renders interpretation difficult – as suggested by the conserved nature of this region among virus isolates.

Conclusion:
This study illustrates the importance to consider the entire sequence of the V3 loop for tropism determination. On the other side our data clearly show the limited contribution from mutations and variation in the V3-flank to a given viral tropism.
Of note, the biggest changes in the FPR and thereby shift towards a tropism change was seen for isolates of the –AE subtype CRF_01.
Larger data sets are still needed to gain a better validity for clinical utility of our findings.