Report

Quantitative Image Analysis of HIV-1 Infection in Lymphoid Tissue

See allHide authors and affiliations

Science  08 Nov 1996:
Vol. 274, Issue 5289, pp. 985-989
DOI: 10.1126/science.274.5289.985

Figures

  • Fig. 1.

    Quantitative image analysis of HIV RNA in LT. (A) A MNC with HIV RNA (indicated by the arrow) in an autoradiograph of a section of tonsil (bright-field and epipolarized illumination, original magnification, ×400). The specular reflectance from the silver grains imparts a yellow-green color to the grains. (B) High-contrast black and white image of the same field illuminated with epipolarized light. Only silver grains and other highly reflective objects in the section relevant to the analysis stand out. (C) For computer-assisted analysis, the image is digitized into 512 by 480 pixels. The area (in squared pixels) occupied by the silver grains in the image is measured after the grains are marked with the MetaMorph threshold tool, which overlays objects in red whose gray values are less than the threshold value set with the tool. The tracing tool is used to delineate the region of silver grains overlying the infected cell that will be measured (white trace). (D) Black and white video image of HIV RNA in virions associated with FDCs in a GC in a section of tonsil (epipolarized illumination). To include the entire GC, the image was captured at ×160. Silver grains were counted at ×600 in segments of the GC (delineated by the white lines) and summed to accurately determine the number in the entire GC. (E) Silver grains in the GC highlighted by a red overlay from the MetaMorph threshold tool.

  • Fig. 2.

    Determination of relevant LT areas by image analysis. (A) The MetaMorph tracing tool was used to trace (in red) a submucosal region of LT in a section of a tonsillar biopsy (original magnification, ×20). The area in pixels was converted to area in millimeters by applying a calibration relating pixel area to square millimeters from a ×20 image of a 2-mm scale with 0.1-mm divisions. (B) An area of white pulp in a section of spleen. The more darkly stained white pulp was identified and traced in color with the MetaMorph autotrace tool. The individual areas were summed to determine total area.

Tables

  • Table 1.

    In situ hybridization and quantitative image analysis of HIV-1 RNA levels in LT. LTs were obtained from HIV-1-infected individuals in the Centers for Disease Control and Prevention (CDC) clinical stage shown and with the CD4 counts indicated. Tissues were fixed immediately to eliminate autolysis as a variable by immersion for 4 to 5 hours in 4% paraformaldehyde, or for at least 24 hours in Streck's tissue fixative, followed by transfer to 70 to 80% ethanol and paraffin embedding. In preliminary experiments, these procedures for tissue processing were determined to be optimal and comparable for detection of HIV RNA and preservation of cellular morphology. The number of copies of HIV-1 RNA was determined by in situ hybridization and quantitative image analysis. HIV-1 RNA levels per milliliter of plasma were determined by branched DNA (bDNA) assay (22) in samples taken at the time of tonsil biopsy (patients 4 through 7 were biopsied several times as indicated). The viral clearance rate, cV, which under quasi-steady-state conditions should equal viral production, was calculated assuming c = 3 day−1 (19) and that V is the total viral load in extracellular fluid calculated by dividing the number of HIV RNA equivalents per milliliter of plasma by 2 and then multiplying by the 15 × 103 ml of extracellular fluid in a 70-kg individual. The rate of virus production in LT, NδT*, was calculated assuming that N is about equal to the maximum of the range of the number of copies of HIV RNA per MNC divided by 2. T* was taken to be the frequency of productively infected MNCs per gram of tissue multiplied by the assumed 700 g of LT per 70-kg individual, and δ = 0.5 day−1 (19). N, T*, and c will be underestimates because, respectively, only about 75% of a productively infected MNC will be analyzed in an 8-μm section, because T* does not include infected cells with fewer than 20 copies of viral RNA, and because of the methods used to estimate c (19). 3TC, Lamivudine-Epivir; D4T, Stavudine-Zerit; DDC, Zalcitabine-HIVID; DLV, Delavirdine; ZDV, Zidovudine-Retrovir. Dashes indicate not applicable.

    CDC classTreatmentCD4 count/mm3 of bloodSampleNumber of HIV RNA equivalents/ml of plasma (bDNA assay)Number of copies of HIV RNA/g of tissueAverage number (and range) of copies of HIV RNA per MNCFrequency of productively infected MNCs/g of tissuecV/NδT*
    FDCsMNCs
    Patient number 1 (age 26 years)
    B3ZDV + 3TC114Tonsil, right4 × 1042.3 × 1082.4 × 10662 (20–188)3.8 × 1040.7
    ZDV + 3TC-Tonsil, left-4.9 × 1083.9 × 10686 (30–259)4.5 × 104
    Patient number 2 (age 32 years)
    A2ZDV + DLV375Tonsil, right<5 × 1031.6 × 1082.9 × 10673 (26–130)4 × 104
    ZDV + DLV-Tonsil, left-9.2 × 1071.9 × 10658 (36–118)3.3 × 104
    Patient number 3 (age 31 years)
    A1ZDV638Tonsil<5 × 1033.2 × 1088.6 × 10562 (23–140)1.4 × 104
    Patient number 4 (age 45 years)
    A1None551Tonsil, baseline1041.9 × 1081.6 × 10676 (28–214)2 × 1040.3
    None-Tonsil, +1 month-1.6 × 1081.3 × 10579 (24–234)1.6 × 103
    None763Tonsil, +6 months<5 × 1035.4 × 1084.6 × 105150 (20–287)3 × 103
    None-Tonsil, +12 months-4.3 × 1081.4 × 10779 (21–188)1.8 × 105
    None437Tonsil, +14 months1.4 × 1043.4 × 1084 × 10694 (20–180)4.2 × 1040.2
    Patient number 5 (age 32 years)
    A2ZDV294Tonsil, baseline<5 × 1031088 × 10480 (20–97)103
    ZDV-Tonsil, +1 month-6.2 × 1078 × 10480 (20–108)103
    ZDV + DLV215Tonsil, +6 months<5 × 1036.3 × 1077.5 × 10475 (49–91)103
    None202Tonsil, +12 months<5 × 1037 × 1073 × 10658 (20–151)5.1 × 104
    None197Tonsil, +14 months<5 × 1035 × 1073.4 × 10657 (20–106)6 × 104
    Patient number 6 (age 32 years)
    A1None633Tonsil, baseline7.6 × 1042.3 × 1073.9 × 10638 (20–141)1050.7
    None-Tonsil, +1 month-1.3 × 1072.2 × 10672 (20–177)3 × 104
    ZDV + DDC389Tonsil, +12 months<5 × 1035.8 × 1071.7 × 10641 (20–115)4 × 104
    Patient number 7 (age 48 years)
    A1ZDV782Tonsil, baseline2.6 × 1044.2 × 1061.8 × 106124 (48–222)1.5 × 1041
    ZDV-Tonsil, +1 month-7.6 × 1071.9 × 10774 (20–183)2.6 × 105
    ZDV778Tonsil, +6 months2.9 × 1041.8 × 1086 × 10585 (50–188)7.1 × 1032.8
    ZDV615Tonsil, +14 months7.6 × 1043.2 × 10710744 (20–136)2.3 × 1050.3
    Patient number 8 (age 37 years)
    A3D4T140Spleen<5 × 1037.9 × 1061.8 × 10558 (45–77)3.1 × 103
    Patient number 9 (age 39 years)
    A2DDC245Lymph node<5 × 1033.8 × 1071.1 × 10663 (28–124)1.7 × 104
    None106Spleen2 × 1057 × 1075.7 × 10663 (20–128)9 × 1042.2
    Mean4021.5 × 1083.4 × 106745.3 × 1041.0
  • Table 2.

    Comparison of single-cell and population estimates of HIV-1 RNA. Simultaneous biopsies of right and left palatine tonsil were obtained and divided into two weighed fragments. One portion was snap frozen, and at a later time RNA was extracted and the number of HIV RNA equivalents determined by bDNA assay (22). The other portion was fixed and sectioned at a later time, and viral RNA associated with FDCs was quantitated as described in this report.

    CDC classTreatmentCD4 count/mm3 of bloodSampleNumber of copies of HIV RNA associated with FDCs/g of tonsillar tissue
    In situ hybridizationbDNA assay
    Patient number 1 (age 26 years)
    B3ZDV + 3TC110Right tonsil2.3 × 1083.9 × 108
    ZDV + 3TC110Left tonsil4.9 × 1085.2 × 108
    Patient number 2 (age 32 years)
    A2ZDV + DLV375Right tonsil1.6 × 1088.1 × 107
    ZDV + DLV375Left tonsil9.2 × 1078.5 × 107