Induction of apoptosis of metastatic mammary carcinoma cells in vivo by disruption of tumor cell surface CD44 function

Abstract

To understand how the hyaluronan receptor CD44 regulates tumor metastasis, the murine mammary carcinoma TA3/St, which constitutively expresses cell surface CD44, was transfected with cDNAs encoding soluble isoforms of CD44 and the transfectants (TA3sCD44) were compared with parental cells (transfected with expression vector only) for growth in vivo and in vitro. Local release of soluble CD44 by the transfectants inhibited the ability of endogenous cell surface CD44 to bind and internalize hyaluronan and to mediate TA3 cell invasion of hyaluronan-producing cell monolayers. Mice intravenously injected with parental TA3/St cells developed massive pulmonary metastases within 21-28 d, whereas animals injected with TA3sCD44 cells developed few or no tumors. Tracing of labeled parental and transfectant tumor cells revealed that both cell types initially adhered to pulmonary endothelium and penetrated the interstitial stroma. However, although parental cells were dividing and forming clusters within lung tissue 48 h following injection, >80% of TA3sCD44 cells underwent apoptosis. Although sCD44 transfectants displayed a marked reduction in their ability to internalize and degrade hyaluronan, they elicited abundant local hyaluronan production within invaded lung tissue, comparable to that induced by parental cells. These observations provide direct evidence that cell surface CD44 function promotes tumor cell survival in invaded tissue and that its suppression can induce apoptosis of the invading tumor cells, possibly as a result of impairing their ability to penetrate the host tissue hyaluronan barrier.

Figure 1

Binding of HA and CD44 expression by TA3/St cells. (A) Binding of FL-HA by TA3/ST cells as assessed by FACS^® analysis (a). The ability of TA3/St cells to bind FL-HA completely abrogated by the blocking anti-CD44 monoclonal antibody KM201 (b), whereas control anti–ICAM-1 mAb has no effect (c). (B) Expression of multiple CD44 isoforms by TA3/St cells: RT-PCR was performed using the oligonucleotide primers listed in Materials and Methods and total RNA from TA3/St cells, and the products were analyzed on 1% agarose gels. Lane 1, 100-bp DNA reference ladder (GIBCO BRL); lane 2, PCR products using a forward primer corresponding to exon 5 (5f) and a reverse primer corresponding to exon 16 (16r). A 120-bp product representing the expression of the standard CD44 isoform containing no variant exons (CD44H) is indicated (white arrow); the presence of larger products indicates expression of multiple CD44 variants. Lanes 3–12, the reverse 16r primer was used together with forward primers v1f through v10f. The products in these lanes demonstrate that TA3/St cells express a range of CD44 variants containing variant exons v1 through v10. (C) Western blot analysis of CD44 expression in lysates (a) and supernatants (b) of parental and soluble CD44-transfected TA3/St cells. Lysates in a and supernatants in b were from: lane 1, TA3neo No. 1; lane 2, TA3neo No. 8; lane 3, TA3sCD44 v8-v10 No. 13; lane 4, TA3sCD44v8-10 No. 19; lane 5, TA3sCD44v6-v10 No. 12; lane 6, TA3sCD44v6-10 No. 17 cells. TA3neo No. 1 and No. 8 cells express CD44H (∼80 kD), and several larger CD44 isoforms; lysates from TA3 cells expressing soluble CD44 show a similar pattern of CD44 isoform expression, except that the abundance of CD44 proteins of ∼80 Kd, which corresponds to the M_r of soluble CD44 isoforms, is increased as expected. Western blot analysis of TA3/St cell culture supernatants (b) reveals that TA3neo No. 1 and No. 8 cells (lanes 1 and 2) do not produce soluble CD44, whereas TA3sCD44 transfectants produce variable amounts of soluble CD44 (lanes 3–6). Arrows indicate molecular weight markers, which are, from top to bottom, 203, 118, 86, and 52 kD.

Figure 1

Binding of HA and CD44 expression by TA3/St cells. (A) Binding of FL-HA by TA3/ST cells as assessed by FACS^® analysis (a). The ability of TA3/St cells to bind FL-HA completely abrogated by the blocking anti-CD44 monoclonal antibody KM201 (b), whereas control anti–ICAM-1 mAb has no effect (c). (B) Expression of multiple CD44 isoforms by TA3/St cells: RT-PCR was performed using the oligonucleotide primers listed in Materials and Methods and total RNA from TA3/St cells, and the products were analyzed on 1% agarose gels. Lane 1, 100-bp DNA reference ladder (GIBCO BRL); lane 2, PCR products using a forward primer corresponding to exon 5 (5f) and a reverse primer corresponding to exon 16 (16r). A 120-bp product representing the expression of the standard CD44 isoform containing no variant exons (CD44H) is indicated (white arrow); the presence of larger products indicates expression of multiple CD44 variants. Lanes 3–12, the reverse 16r primer was used together with forward primers v1f through v10f. The products in these lanes demonstrate that TA3/St cells express a range of CD44 variants containing variant exons v1 through v10. (C) Western blot analysis of CD44 expression in lysates (a) and supernatants (b) of parental and soluble CD44-transfected TA3/St cells. Lysates in a and supernatants in b were from: lane 1, TA3neo No. 1; lane 2, TA3neo No. 8; lane 3, TA3sCD44 v8-v10 No. 13; lane 4, TA3sCD44v8-10 No. 19; lane 5, TA3sCD44v6-v10 No. 12; lane 6, TA3sCD44v6-10 No. 17 cells. TA3neo No. 1 and No. 8 cells express CD44H (∼80 kD), and several larger CD44 isoforms; lysates from TA3 cells expressing soluble CD44 show a similar pattern of CD44 isoform expression, except that the abundance of CD44 proteins of ∼80 Kd, which corresponds to the M_r of soluble CD44 isoforms, is increased as expected. Western blot analysis of TA3/St cell culture supernatants (b) reveals that TA3neo No. 1 and No. 8 cells (lanes 1 and 2) do not produce soluble CD44, whereas TA3sCD44 transfectants produce variable amounts of soluble CD44 (lanes 3–6). Arrows indicate molecular weight markers, which are, from top to bottom, 203, 118, 86, and 52 kD.

Figure 1

Binding of HA and CD44 expression by TA3/St cells. (A) Binding of FL-HA by TA3/ST cells as assessed by FACS^® analysis (a). The ability of TA3/St cells to bind FL-HA completely abrogated by the blocking anti-CD44 monoclonal antibody KM201 (b), whereas control anti–ICAM-1 mAb has no effect (c). (B) Expression of multiple CD44 isoforms by TA3/St cells: RT-PCR was performed using the oligonucleotide primers listed in Materials and Methods and total RNA from TA3/St cells, and the products were analyzed on 1% agarose gels. Lane 1, 100-bp DNA reference ladder (GIBCO BRL); lane 2, PCR products using a forward primer corresponding to exon 5 (5f) and a reverse primer corresponding to exon 16 (16r). A 120-bp product representing the expression of the standard CD44 isoform containing no variant exons (CD44H) is indicated (white arrow); the presence of larger products indicates expression of multiple CD44 variants. Lanes 3–12, the reverse 16r primer was used together with forward primers v1f through v10f. The products in these lanes demonstrate that TA3/St cells express a range of CD44 variants containing variant exons v1 through v10. (C) Western blot analysis of CD44 expression in lysates (a) and supernatants (b) of parental and soluble CD44-transfected TA3/St cells. Lysates in a and supernatants in b were from: lane 1, TA3neo No. 1; lane 2, TA3neo No. 8; lane 3, TA3sCD44 v8-v10 No. 13; lane 4, TA3sCD44v8-10 No. 19; lane 5, TA3sCD44v6-v10 No. 12; lane 6, TA3sCD44v6-10 No. 17 cells. TA3neo No. 1 and No. 8 cells express CD44H (∼80 kD), and several larger CD44 isoforms; lysates from TA3 cells expressing soluble CD44 show a similar pattern of CD44 isoform expression, except that the abundance of CD44 proteins of ∼80 Kd, which corresponds to the M_r of soluble CD44 isoforms, is increased as expected. Western blot analysis of TA3/St cell culture supernatants (b) reveals that TA3neo No. 1 and No. 8 cells (lanes 1 and 2) do not produce soluble CD44, whereas TA3sCD44 transfectants produce variable amounts of soluble CD44 (lanes 3–6). Arrows indicate molecular weight markers, which are, from top to bottom, 203, 118, 86, and 52 kD.

Figure 2

(A) Expression of soluble CD44 prevents lung metastasis of TA3 cells. Lungs from two representative mice injected with the TA3sCD44v6-10 No. 17 cells reveal no apparent tumor nodules (a and b); lungs from two representative mice injected with TA3neo No. 1 display innumerable tumor nodules throughout the lung parenchyma (c and d). (B) Tumors derived from lungs of mice injected with TA3sCD44v6-v10 No. 12 cells have lost soluble CD44 expression. Five independent tumor nodules were dissected out from the lung, and RT-PCR was performed using total RNA derived from each nodule (lanes 2–11). The primers used were: lanes 2, 4, 6, 8, 10, and 14, primer 1f and primer 20r, to detect expression of endogenous CD44; lanes 3, 5, 7, 9, 11, and 15, primer 1f and new v10r to detect expression of soluble CD44 (16); lanes 12–15: total RNA from cultured TA3sCD44v6-v10 No. 12 cells was used as template for lanes 12 and 13, negative controls, where PCR was performed with 1f primers only; lanes 14 and 15, positive controls where PCR was performed with primers 1f and 20r (lane 14), and 1f and new v10r (lane 15). Molecular weight markers (100-bp ladder) are shown in lane 1. RT-PCR analysis indicates that all five tumor nodules express endogenous but not soluble recombinant CD44.

Figure 3

(A) Soluble CD44 inhibits events required for TA3/St cell growth in the lung tissue microenvironment. TA3neo No. 1 (a, c, and e) and TA3sCD44v6-10 No. 17 (b, d, and f) cells were labeled with green CMFDA and 5 × 10⁶ cells in 0.2 ml HBSS buffer were injected into the tail vein of A/jax mice. The animals were killed at 1, 24, and 48 h after injection, the lungs were fixed and paraffin-embedded, and 5-μm-thick paraffin sections were mounted onto slides and examined by fluorescence microscopy. 1 h after injection, both TA3neo No. 1 and TA3sCD44 v6-v10 No. 17 cells were observed to be arrested in pulmonary blood vessels (a and b) and to penetrate the pulmonary interstitium (c and d). Occasional cells appeared to be in the process of extravasation (c, arrowhead and the arrow indicate an extravasating cell and the lumen of a blood vessel, respectively). 48 h after injection (e and f), TA3 neo cells display cluster formation (e), whereas only a few isolated TA3sCD44v6-10 No. 17 cells remain in the lung parenchyma (f). Bar in a and b = 304 μm; bar in c–f = 76 μm. (B) In situ detection of apoptosis by TUNEL assay. Lung sections showing TA3neo No. 1 cells (a and c) and TA3sCD44v6-10 No. 17 cells (b and d) labeled with CMFDA (green fluorescence) or detected with ApopTag (red fluorescence) 1 h (a and b) and 48 h (c and d) after intravenous injection. 1 h after injection, TA3neo No. 1 and TA3sCD44 v6-v10 No. 17 cells display similar distribution and no reactivity with ApopTAG (red fluorescence). 48 h after injection, the majority of TA3neo cells were ApopTAG-negative (c) whereas most (>80%) of the remaining TA3sCD44v6-v10 No. 17 cells tested positive for ApopTAG staining (d, arrows). Bar in a–d = 125 μm.

Figure 4

(A) [³H]-HA uptake and degradation. Soluble CD44 expression by the transfected TA3 cells results in a reduction of up to three-fold of ³H-HA uptake and degradation. (B) Invasion of G8 myoblast monolayers by TA3neo and TA3sCD44 cells. TA3neo (a and c) and TA3sCD44v6-10 (b and d) cells (5 × 10³ cells/well) were seeded onto DMSO-fixed G8 monolayers which were untreated (a and b) or treated with streptomyces hyaluronidase (10 U/ml, ICN, Costa Mesa, CA; c and d) at 37°C for 3 h. 6 d after seeding of TA3 transfectants, TA3 neo cells are observed to penetrate and invade the monolayer, treated or not with hyaluronidase, and form tumor cell islets (arrows, TA3 neo cells). By contrast, TA3sCD44 cells were unable to penetrate the untreated G8 monolayer (b, arrowheads), whereas treatment of the fixed G8 cells with hyaluronidase restored the ability of TA3sCD44 cell to penetrate the monolayer (d, open arrows).

Figure 4

(A) [³H]-HA uptake and degradation. Soluble CD44 expression by the transfected TA3 cells results in a reduction of up to three-fold of ³H-HA uptake and degradation. (B) Invasion of G8 myoblast monolayers by TA3neo and TA3sCD44 cells. TA3neo (a and c) and TA3sCD44v6-10 (b and d) cells (5 × 10³ cells/well) were seeded onto DMSO-fixed G8 monolayers which were untreated (a and b) or treated with streptomyces hyaluronidase (10 U/ml, ICN, Costa Mesa, CA; c and d) at 37°C for 3 h. 6 d after seeding of TA3 transfectants, TA3 neo cells are observed to penetrate and invade the monolayer, treated or not with hyaluronidase, and form tumor cell islets (arrows, TA3 neo cells). By contrast, TA3sCD44 cells were unable to penetrate the untreated G8 monolayer (b, arrowheads), whereas treatment of the fixed G8 cells with hyaluronidase restored the ability of TA3sCD44 cell to penetrate the monolayer (d, open arrows).

Figure 5

HA production at sites of tumor invasion. (a–d) Lung sections 1 h after tail vein injection of TA3neo No. 1 (a and c) and TA3sCD44v6-10 No. 17 (b and d) reveal no detectable HA production as assessed by staining with biotinylated HA-binding bPG. Arrows indicate invading tumor cells that display large hyperchromatic nuclei. 48 h after tail vein injection (e–h), both TA3 neo No. 1 (e and g) and TA3sCD44v6-10 No. 17 (f and h) induce stromal HA production. TA3neo No. 1 cells are visible at the sites of HA production (g, arrow) whereas TA3sCD44v6-10 No. 17 cells are no longer distinguishable (h, arrowhead).