PATIENTS AND SAMPLES
Papers I, II, III and IV were based on samples collected prospectively between March 1987 and September 1994 following a national program for pregnant women living in Sweden (Lindgren et al., 1993). These samples were all retrieved from the Department of Clinical Virology at Karolinska University Hospital Huddinge at one time point in 1997. All transferred samples have been listed. The lists have been signed by the then chief of the Department of clinical virology, Professor Anders Vahlne (co-signed by Dr Anders Sönnerborg), and the receipt by the then prefect of MTC,
Professor Ingemar Ernberg (co-signed by Professor Eva Maria Fenyö). Subsequently, they have been reported to a biobank register, as a sample collection. Originally they were received and handled at the Central Microbiological laboratory of the Stockholm County Council (SMCL) until June 1994, from which time they were received at Karolinska University Hospital Huddinge. The patients were HIV-1 infected pregnant women and children, most of whom were born to these mothers in Sweden. Samples were collected at birth, and at 6 months intervals, except during the first year when samples were collected also at 3 months, and if born after 1991, also at 6 weeks of age.
Cord blood was not used. After 18 months of age, samples were taken every following half or whole year.
Defibrinated whole blood was used for all virus culture from plasma and PBMC, as well as for HIV-1 detection by the polymerase chain reaction (PCR). The blood had been collected and separated by Ficoll and frozen as plasma supernatants and PBMC pellets. In the Swedish cohort, 220 distinct isolates were analyzed from 39 subjects in papers I to IV. Table 3 shows how they were included in each study, some of which had been described in the previous works. In the summary line, only he unique and non-overlaping numbers are shown. Paper V involved submitted and publicly available sequence data from the International HIV database in Los Alamos and included 176 HIV-1 infected individuals in whom the data included virus, classified according to phenotype and genotype (table 3). Thus, a total of 215 subjects were studies in papers I through V.
Table 3. Overview of the patients and samples included in the papers within this thesis and the total number of sequences and clones involved. A total of 220 isolates were characterized in papers I to IV.
Paper Subjects Isolates Sequences clones
PBMC Plasma
I 24 children 86 57 -
II 11 mother-child pairs 79 59 -
III 2 mother-child pairs 12 8 195
IV 5 mothers 17 12 290
V 176 individuals - - 176
In all 215 129 91 620
METHODS
Virus Isolation (Papers I to IV)
Virus isolates were co-cultured with phytohemagglutinin stimulated PBMC from two healthy blood donors (Ehrnst et al., 1988). Virus stocks were obtained through passage through donor PBMC and monitored through gag p24 antigen ELISA (papers I, II and III) or through env V3 PCR (paper IV).
Determination of Co-receptor Use (Papers I to IV)
Co-receptor use was determined by inoculating U87 astroglioma cell lines expressing CD4 and chemokine receptors CCR5 or CXCR4 (paper IV) and other co-receptors CCR1, CCR2b, CCR3 (Deng et al., 1996; Deng et al., 1997; Berger et al., 1998). Additional co-receptor use determination was performed on GHOST(3) cells expressing CCR5, CXCR4 or orphan receptors BOB or BONZO (Deng et al., 1997) (papers I, II and III).
Sample Preparation for PCR (Papers I to IV)
Prior to PCR assays, infected PBMC or U87 cells were treated with a lysis buffer, containing Proteinase-K at different temperatures to first enable the proteinase-K to break down the cellular structures and expose the cell DNA and later to inactivate both virus and the active enzyme, rendering it safe for regular laboratory work and of good quality for PCR.
Nested PCR (Papers I to IV)
A nested PCR of the V3 region of gp120 was used as a base for subtyping of HIV-1, to create molecular clones for DNA sequencing of isolates and clones, and lately for surveillance of HIV-1 infection in PBMC cultures and in co-receptor use
determinations.
Generation of DNA Clones (Papers III and IV)
To amplify integrated proviral DNA from one single cell, a limiting dilution technique based on the Poisson theory of random distribution was used (Brinchmann et al., 1991; Leitner et al., 1996a; Rodrigo et al., 1997). Quadruplet sets of dilutions for each sample were set up with an increasing four-fold dilution. The nested PCR
described above was used and an agarose gel electrophoresis conducted on all samples.
An average dilution was extrapolated from the resulting electrophoresis which would render one third of the reactions positive, likely representative of one provirus. Based on this dilution, a second round of nested PCR was performed on each sample with approximately 50 reactions. This give on average 20 positive clones, confirmed using agarose gel electrophoresis. As it was still difficult to accurately predict the outcome of positive reactions, we introduced a practice of using two sets of dilutions from which the best set was chosen for sequencing.
Prior to sequencing, PCR samples were purified using nitrocellulose-membrane based filtration techniques to remove salts and other left over PCR components.
DNA Sequencing (Papers I to IV)
Sequencing PCR was performed on all positive PCR reactions from both isolate and clone samples, subjected to nested PCR. This cycle sequencing PCR was conducted using both inner primers JA168 and JA169 in separate tubes to provide two complementary sequences from each sample, enabling optimal base-calling (Leitner et al., 1996b). Samples were afterwards purified with
nitrocellulose-membrane based filtration techniques specific for sequencing reaction clean-ups.
Sequence Processing (Papers I to IV)
Raw sequence data was analyzed using base-calling software to determine an accurate consensus nucleotide sequence for each sample based on both
complementary sequences derived through each inner primer. This provided a tool for scrutinizing ambiguities and possible mismatches in the raw data
chromatograms and using correct IUPAC letter code abbreviation to identify true ambiguities. In general, multiple chromatogram peaks present on both
complementary strands with a height of at least 25% of the largest peak were accepted as ambiguities and given their appropriate IUPAC code. The criteria set up for accepting clone sequences for inclusion in our analysis were a maximum of one ambiguity. This was to ensure that each sequence we obtained was accurate in its nucleotide composition and when an ambiguity was found, present this sequence in its two configurations with designated suffixes a, and b. Sequences with multiple ambiguities were omitted as the original PCR amplification probably represented multiple pro-viruses and not a single molecular clone. The isolate sequences were allowed to contain multiple ambiguities. Prior to analysis, base-called sequenced were assembled with Se-Al sequence alignment software (Rambaut, 1995).
Hypermutation Analysis (Papers III)
Hypermutations were assessed among HIV-1 sequences in paper III using
Hypermut (Rose & Korber, 2000) with maternal sequence M1.2aD as master sequence (#GÆA>10, normal value 0-5). Sequences with abnormal GÆA mutations (Vartanian et al., 1991; Rose & Korber, 2000) were omitted in the analyses of the mother-to-child transmission to facilitate maximum resolution in our phylogenetic trees.
Phylogenetic Analysis (Papers III and IV)
An important tool in analyzing the data presented in this thesis has been the use of phylogeny. It is a tool for studying the evolutionary relatedness of species, or HIV-1 sequences in our cases, and provides a schematic overview of the ancestral
relationships between our clones and isolates, referred to as a phylogenetic tree. It is also a useful tool for studying transmission (Leitnet et al., 1996b; Albert et al., 1994).
Phylogenetic tree analysis was used to assess subtypes and virus evolution over time in papers I to IV. For subtyping of all isolates, a preliminary amino acid pattern analysis was performed where the sequences in question was matched to a pool of subtype specific consensus sequences, providing a percentage value of association to its correct clade. The genetic subtype was later confirmed by phylogenetic tree analysis using the programs DNADIST and NEIGHBOR in the PHYLIP package (Felsenstein, 1993) and recommended reference sequences from the Los Alamos Database. The phylogenetic analysis in paper IV was also analyzed using DNADIST and NEIGHBOR through the online TREEMAKER tool from the HIV Sequence Database in Los
Alamos, NM.
For paper III, phylogenetic analysis for transmission and X4 evolution analysis was conducted using DNAML, a maximum likelihood method, to accurately
reconstruct the transmission histories of the two mother-child pairs using the BioEdit program (Hall, 1999). First, a maximum likelihood tree was constructed under the F84 substitution model using a parallel computing system (Felsenstein, 1993 & Leitner &
Albert, 1999). Second, the branch lengths were recalculated under a general-time-reversible model with Г-distributed rates among sites (α=0.3) (Swofford, 2002).
Finally, each tree was rooted to early maternal sequences to provide the trees with a logical point of reference. In addition, a bootstrap analysis was performed to evaluate out tree topology using the neighbor-joining method under the F84 model with 1000 resamplings (Felsenstein, 1993).
Statistical Analysis (Papers I to V)
Statistical methods were used in all papers. The non-parametric Fischer’s exact test, (and McNemar´s test), and the binomial test were used to assess statistical probability in our data. In paper V, sensitivity and specificity values were calculated for the amino acid motifs, as well as their predictive values. Herein, the binomial test was employed in a further analysis of the data in paper II.
Ethical Aspects
Ethical permission for the studies presented in this thesis was given by the Ethics Committee of the Karolinska Institute and consent was given by the HIV-1 infected mothers.
Notes on Patient Codes and Sample Processing
During handling and processing of the samples from the point of collection to the moment of laboratory work for the presented studies, multiple safeguards have been used to reduce the risk of sample mix-ups and contamination.
Safety rules include both protection of patient integrity and safety of the laboratory worker from becoming infected. Patients were assigned a study number where the mother was the index case and her children given letters related to being firstborn, etc.
No personal data from the patients were available in the laboratory at MTC. Thus all patient data were pseudonymized. This is a terminology introduced for bio-banks, when there is a coupling to the patient code somewhere, but not where the samples are dealt with. In this case the patient identities were available in the Department of Clinical Virology at the Karolinska University Hospital Huddinge and among the clinicians involved. Dr Susanne Lindgren was the principle investigator of gynecological issues, and Dr Ann-Britt Bohlin for the children.
For each time point there was a protocol number, running continuously from start.
In addition, blood components, such as plasma, serum, PBMC for isolation or for PCR were given individual sample numbers running through-out a year with the year as prefix. Each sample was labeled with the date of handling, in addition to the protocol and sample numbers. Mothers and children had separate protocol numbers, even when arrived in the same parcel. Thus it was possible to make an adequate guess about what number to assume, when a handwritten label on a tube was difficult to interpret.
Whenever a sample was thawed and refrozen this was marked on the tube. When larger volumes were thawed for the first time, they were re-aliquoted into smaller volumes or into a combination of larger and smaller volumes in order to keep the number of freeze-thawing to a minimum.
The samples were kept in a freezer at minus 85˚C. As they were retrieved from a larger set of samples, the numbers were not continuous, which made identification time-consuming. As there also was a hazard of dealing with samples, containing infectious HIV, the safety-guard from this point of view, required that we were two people when picking out samples. One person kept track of the list of samples, the other in laboratory clothes, adequate for the situation, picked out the samples. It became necessary to fill empty spots with paper. Otherwise it was easy that cold samples
attached to the glove and then found its way into a new spot.
Our own protocols were always set up in sequence of the protocol number on the tubes. This usually meant that individual samples from the same patients were not run directly together. Again this facilitated interpretation of handwritten figures. We also had protocol numbers with another letter than the original protocol number as prefix.