NWCS 261

1
Association of HIV-1 Co-receptor Tropism with
Immunologic and Virologic Parameters in HIV-1
infected, Treatment-Naïve Subjects in ACTG 384 G
Skowron1, E Chan2, J Weidler3, G Robbins4, V
Johnson5, J Spritzler2, D Asmuth6, R Gandhi4, Y
Lie3, E Coakley3, R Pollard6, for the AIDS
Clinical Trials Group 384 protocol team. 1
Roger Williams Medical Center/The Miriam Hospital
ACTU, Providence, RI, and Boston Univ School of
Medicine, Boston, MA, USA, 2 Harvard School of
Public Health, Boston, MA, USA, 3 Monogram
Biosciences, Inc., South San Francisco, CA, USA,
4 Massachusetts General Hospital, Boston, CA,
USA, 5 Birmingham VA Medical Center and
University of Alabama at Birmingham School of
Medicine, Birmingham, AL, USA, 6 University of
California Davis Medical School, Sacramento, CA,
USA
Gail Skowron, M.D. Division of Infectious
Diseases Roger Williams Medical Center 825
Chalkstone Ave, Providence, RI 02908 gail_skowron_at_
brown.edu telephone [phone] FAX
401-456-6839
Poster 388 CROI 2009 8-11 Feb Montreal, Canada
Results
Background
Discussion
Viral co-receptor tropism is associated with
disease progression in untreated chronic HIV-1
infection. The relationships between tropism and
other pre-treatment factors, such as replication
capacity and lymphocyte subsets, have not been
described. Higher baseline (BL) viral replication
capacity (RC) has previously been associated with
higher HIV-1 RNA and CD4 activation , and lower
CD4 and CD4 memory counts.
Figure 1. Baseline Immunologic and Virologic
Parameters by Baseline Tropism
Table 3. Linear Regression Model of Baseline CD8
activation percent
1. In this analysis of selected treatment-naive
subjects participating in ACTG 384, BL DM/X4
tropism was associated with lower BL memory,
naive and total CD4 count, higher BL HIV-1 RNA,
higher replication capacity, and higher CD4 and
CD8 percent activation (Table 2). 2. We
previously demonstrated, in an analysis of a
larger subset of subjects participating in ACTG
384, that higher baseline RC was significantly
associated with higher HIV-1 RNA, higher CD4 cell
activation, lower CD4 cell count, and lower CD4
memory cell count (Skowron, 2009). 3. In a
linear regression model (Table 3), higher
baseline CD8 activation percent was significantly
associated with higher baseline viral load (RNA),
controlling for tropism, CD4 cell count and
replication capacity at baseline. 4. In a linear
regression model (Table 4), higher baseline CD4
activation percent was significantly associated
with non-R5 tropism, lower baseline CD4 cell
count, and higher baseline replication capacity
(RC), adjusted for BL RNA. We had previously
noted an association between higher baseline RC
and higher baseline CD4 activation percent (r
0.23, p lt 0.0001). 5. In our previous analysis,
we constructed a multivariable model,
demonstrating that an increase in CD4 cell count
from baseline to week 48 was associated with
lower baseline CD4 memory count and higher
baseline CD4 naive percent (p0.004, p0.015,
respectively), adjusting for baseline RC,
baseline CD4 and CD8 count, baseline RNA, and the
interaction between baseline CD4 and baseline RC.
The current analysis adds baseline tropism to
this model (Table 5), but only CD4 naive percent
remains significant, i.e., an increase in CD4
cell count is associated with higher baseline CD4
naive percent, controlling for the above baseline
covariates and tropism. 6. These data raise
provocative questions regarding the interplay
between RC and DM/X4 tropism, in depleting naive
and memory CD4 cells and increasing CD4
activation, and the resultant effect on CD4
recovery in patients on effective ART.
Variable Parameter Estimate Standard Error t value Pr gt t
Intercept 36.97 10.55 3.50 0.0006
BL Tropism R5 -1.37 2.75 -0.05 0.62
BL log10 RNA 4.02 1.86 2.15 0.033
BL CD4 cell count -0.010 0.0067 -1.52 0.13
BL RC -0.019 0.020 -0.95 0.34
Methods
Table 4. Linear Regression Model of Baseline CD4
activation percent
Using Monograms original Trofile Assay, we
determined HIV-1 co-receptor tropism on
pre-treatment plasma samples from 230
HIV-infected, treatment-naive subjects enrolled
in ACTG 384 and selected on CD4 and viral
response status (a subset of subjects with sample
available from the group described in Skowron,
2009). Of these, 210 had undetectable HIV-1 RNA
(?50) at week 48. HIV-1 RNA, CD4 and CD8 subsets
and RC were used to investigate BL and week 48
associations with tropism. Continuous outcomes
were compared between R5 vs. DM/X4 groups with
Wilcoxon Rank Sum tests. Linear and logistic
regression models were also applied when
controlling for BL covariates.
Variable Parameter Estimate Standard Error t value Pr gt t
Intercept 30.51 9.19 3.32 0.0011
BL Tropism R5 -9.98 2.39 -4.17 lt 0.0001
BL log10 RNA 0.71 1.62 0.44 0.66
BL CD4 cell count -0.030 0.0058 -5.14 lt 0.0001
BL RC 0.039 0.018 2.19 0.0296

Table 2. Baseline Immunologic and Virologic
Parameters by Baseline Tropism
Table 5. Linear Regression Model of Change in
CD4 cell count from BL to wk 48
BL covariates           Tropism  Median (Q1, Q3) P (Wilcoxon) P (adj. for RC)
BL CD4 Cell Count (cells/mm3)                               R5    385 (220, 525) lt0.0001 lt0.0001
BL CD4 Cell Count (cells/mm3)                               DM or X4  138 (24, 311)  lt0.0001 lt0.0001
BL Memory CD4 Count (cells/mm3) CD4/CD45RO/CD45RA-  R5 192 (116, 271) lt0.0001 0.0001
BL Memory CD4 Count (cells/mm3) CD4/CD45RO/CD45RA-  DM/X4 70 (16, 173)  lt0.0001 0.0001
BL Naive CD4 Count (cells/mm3) CD4/CD45RA/CD62L R5 130 (49, 219) lt0.0001 lt0.0001
BL Naive CD4 Count (cells/mm3) CD4/CD45RA/CD62L DM/X4  19 (3, 99) lt0.0001 lt0.0001
BL CD4 Activation Percent () CD4/CD38/HLA-DR R5    12 ( 7, 19) lt0.0001 lt0.0001
BL CD4 Activation Percent () CD4/CD38/HLA-DR DM/X4  33 (13, 48)    lt0.0001 lt0.0001
BL CD8 Activation Percent () CD8/CD38/HLA-DR R5 48 (37, 60) 0.046 0.047
BL CD8 Activation Percent () CD8/CD38/HLA-DR DM/X4 53 (41, 62)     0.046 0.047
BL log 10 RNA (log10 copies/ml) R5 4.8 (4.3, 5.3) lt0.0001 0.0004
BL log 10 RNA (log10 copies/ml) DM/X4 5.3 (4.9, 5.6) lt0.0001 0.0004
BL Replication Capacity (BL RC) R5 89 (53, 124) 0.0014 N/A
BL Replication Capacity (BL RC) DM/X4 118 (82, 153) 0.0014 N/A
Variable             Parameter Estimate   Standard Error   t value    Pr gt t
Intercept           114.75      111.02       1.03      0.30
BL Tropism R5        11.99      27.65       0.43      0.66
BL RC                -0.43        0.28 -1.55      0.12
BL CD4 cell count   -0.16       0.15   -1.03      0.30
BL CD8 cell count -0.037    0.022    -1.67      0.097
BL log10 RNA         24.63  18.60      1.32      0.19
BL CD4 Memory count  -0.12     0.18     -0.66      0.51
BL CD4 Naive percent  1.42     0.72     1.98      0.049
Bl CD4 x BL RC             0.0018  0.00087  2.02      0.045
Results
Table 1. Baseline Characteristics
N N 230
CD4 count (cells/mm3) median (Q1, Q3) CD4 count (cells/mm3) median (Q1, Q3) 350 (119, 478)
HIV-1 RNA (log10copies/mL) median (Q1, Q3) HIV-1 RNA (log10copies/mL) median (Q1, Q3) 4.9 (4.4, 5.4)
Tropism R5 () 173 (75)
Tropism Dual-Mixed () 56 (24)
Tropism X4 () 1 (lt1)

• References
• Skowron G, Spritzler JG, Weidler J, et al.
  Replication Capacity in Relation to Immunologic
  and Virologic Outcomes in HIV-1 infected,
  Treatment-Naïve Subjects,
• J Acquir Immune Defic Syndr, 2009, in press.

DM and X4 subsets combined in this analysis
(DM/X4)
restricted to 210 subjects who had virologic
suppression (RNA ? 50) at wk 48
This project was supported by SBIR Grant Number
R44AI050321 from the National Institute of
Allergy and Infectious Diseases (NIAID) to
Monogram Biosciences and NIAID grant numbers
AI38855, AI27659, AI38858, AI25879, AI27666. The
ACTG 384 study was also supported in part by
Agouron/Pfizer, Bristol Myers Squibb, and
GlaxoSmithKline.
in a linear regression model adjusting for BL RC