Therapeutic benefit of Salmonella attributed to LPS and TNF-α is exhaustible and dictated by tumor susceptibility

Abstract

The potential of bacteria-mediated tumor therapy (BMTT) is highlighted by more than a century of investigation. Attenuated Salmonella has prevailed as promising therapeutic agents. For BMTT - categorized as an immune therapy - the exact contribution of particular immune reactions to the therapeutic effect remains ambiguous. In addition, one could argue for or against the requirement of bacterial viability and tumor targeting. Herein we evaluate the isolated therapeutic efficacy of purified LPS and TNF-α, which together account for a dominant immunogenic pathway of gram negative bacteria like Salmonella. We show that therapeutic efficacy against CT26 tumors does not require bacterial viability. Analogous to viable Salmonella SL7207, tumor regression by a specific CD8+ T cell response can be induced by purified LPS or recombinant TNF-α (rTNF-α). Conversely, therapeutic effects against RenCa tumors were abrogated upon bacterial avitalization and limited using isolated adjuvants. This argues for an alternative mechanistic explanation for SL7207 against RenCa that depends on viability and persistence. Unable to boost bacterial therapies by co-injection of rTNF-α suggested therapeutic effects along this axis are exhausted by the intrinsic adjuvanticity of bacteria alone. However, the importance of TNF-α for BMTT was highlighted by its support of tumor invasion and colonization in concert with lower infective doses of Salmonella. In consideration, bacterial therapeutic effectiveness along the axis of LPS and TNF-α appears limited, and does not offer the necessary plasticity for different tumors. This emphasizes a need for recombinant strengthening and vehicular exploitation to accommodate potency, plasticity and distinctiveness in BMTT.

Keywords: LPS; TNF alpha; bacteria mediated tumor therapy; cancer immune therapy; salmonella.

Figure 1. Dose-response relationship between SL7207 and therapeutic effects

Tumor-bearing mice were inoculated i.v. with titrated doses of viable SL7207, as indicated. (A) CT26 tumor development profiles. N = 5. (B) RenCa tumor development profiles N = 6. Therapeutic potency increases with higher infection dose. (C) Host body weight profiles. N = 5. Increase in body weight loss with higher infection dose. (D) Levels of TNF-α in serum 1.5 hpi, as determined using ELISA. N = 5. Levels of TNF-α increase with higher infection dose. Displayed are Means ± SEM. Asterisks *, ** and *** denote significance levels of p < 0.05, p < 0.01, or p < 0.001, resp.

Figure 2. Single-dose purified LPS induces CT26 clearance and exhibits transient potency against RenCa

Tumor-bearing mice were inoculated i.v. with 5 × 10⁶ viable SL7207 or 50 μg purified LPS from Salmonella typhosa or E. coli O55:B6, before comparing the therapeutic outcomes. (A) CT26 tumor development profiles. N = 5. Treatments display a comparable kinetic of tumor regression. (B) RenCa tumor development profiles N = 6. Transient therapeutic benefit of systemically inoculated LPS. (C) Host body weight profiles. N = 5. Transient weight loss in response to purified LPS, distinct from progressive weight loss during salmonellosis. (D) Levels of TNF-α in sera at 1.5 hpi, as determined using ELISA. N = 3. Dose-response to LPS. Displayed are Means ± SEM. Asterisks *, ** and *** denote significance levels of p < 0.05, p < 0.01, or p < 0.001, resp.

Figure 3. Analogous to SL7207, purified LPS elicits a tumor-specific CD8⁺ T cell response capable of CT26 rejection

Therapeutic effects of pre-stimulated CD8⁺ or CD4⁺ T cell populations were evaluated on CT26 tumor growth in Rag1^−/− mice. (A) Schematic depiction of reconstitution experiments performed. In essence, 1) durable and complete regression of CT26 tumors through primary therapy (here purified “LPS”) in wild-type mice is a prerequisite for T cell isolation. 2) Upon recovery; 3 months post therapy and antibiotic (ciprofloxacin) treatment [77], 3 × 10⁶ CD4⁺ or CD8⁺ T cells are adoptively transferred to Rag1^−/−mice following negative purification from a pool of splenocytes recovered from the wild-type donor. 3) Concurrent s.c. injection of 5×10⁵ CT26 tumor cells allows for evaluation of tumor growth and -establishment post adoptive transfer. (B) CT26 tumor development in Rag1^−/−mice reconstituted with CD8⁺ T cells from wild-type donors pre-stimulated as indicated in the brackets. N = 6. (C) End-point significance comparison of mean tumor volumes in Rag1^−/−mice upon reconstitution with pre-stimulated CD8⁺ T cells. Reconstitution with LPS stimulated CD8⁺ T cells prevents establishment of CT26 tumors (D) CT26 tumor development in Rag1^−/−mice reconstituted with CD4⁺ T cells from wild-type donors pre-stimulated as indicated in the brackets. N = 6. (E) End-point significance comparison of mean tumor volumes in Rag1^−/−mice upon reconstitution with pre-stimulated CD4⁺ T cells. No significant difference to the CT26 control. The (CT26)-group denotes presence of an established, albeit untreated, CT26 tumor on the T cell donor. Displayed are Means ± SEM. Asterisks *, ** and *** denote significance levels of p < 0.05, p < 0.01, or p < 0.001, resp.

Figure 4. While exclusive to CT26, tumor clearance is equally attainable through recombinant TNF-α as with LPS or Salmonella, and dependent on adaptive immunity

1 μg recombinant murine TNF-α (rTNF-α) was administered i.v. and therapeutic effects in wild-type and Rag^−/−mice were evaluated relative to LPS therapy or BMTT. (A) CT26 tumor development in WT mice represented by close-up photographs of CT26 tumors subjected to treatments as indicated. Similar macroscopic profile of tumor darkening, ulceration and clearance between all therapies. Images displayed are representative of three individual replicates (B) Kinetic of CT26 tumor development in Rag1^−/−mice. N = 5. Tumor regression with all therapies is abrogated in absence of a functional adaptive immune system. (C) RenCa and F1.A11 tumor development profiles in WT mice. N = 5. No significant therapeutic effects induced by rTNF-α. Displayed are Means ± SEM.

Figure 5. Prognostic characteristics of the CT26 tumor are shared among different immunogenic treatments, and less pronounced in RenCa

Tumors from treated or untreated mice were isolated 48 hpi, and subjected to immune histochemical staining. Histological images of stained CT26 (A) or RenCa (B) tumor cross sections. More extensive necrosis formation with treatments of CT26 compared to RenCa. Images displayed are representative of four replicates. “N” indicates greater areas of necrosis. Hypoxia was stained with antibodies against metabolites of pimonidazole-HCl, otherwise administered i.v. 30 mins prior to isolation. Myeloperoxidase (MPO) denotes presence of neutrophilic granulocytes, and Ki67 the extent of proliferative activity. Differential staining was performed on consecutive sections. Scale bar corresponds to 100 μm. Orientation of the tumor cross section as indicated.

Figure 6. While dispensable for CT26 regression, viability of SL7207 is cause of an infectious phenotype, and required for the therapeutic efficacy against RenCa

Tumor-bearing mice were inoculated i.v. with 5 × 10⁶ viable or avitalized Salmonella SL7207, and the therapeutic effects compared. Avitalization was facilitated through heat inactivation (HI) at 60°C or UV irradiation. (A) CT26 tumor development profiles. N = 6. Mean ± SEM. Similar kinetic of regression post treatment using viable and avitalized SL7207. (B) Statistical comparison between mean tumor volumes. No difference at indicated time points. (C) RenCa tumor development profiles. N = 6. Mean ± SEM. In contrast to viable SL7207, avitalization abrogates the retarding effect and causes outgrowth. (D) End-point comparison of mean RenCa tumor volumes. No significant difference between UV irradiated and HI groups. (E) CFU counts in RenCa tumors upon 48 hours of treatment, as indicated. N = 3. Median ± range. Only viable therapy facilitates tumor colonization (i.e. persistence) of SL7207 in the host. (F) Host body weight profile dictated by recovery for avitalized groups, and progressive loss with viable SL7207. N = 6. Mean ± SEM. (G) Levels of TNF-α in sera at 1.5 hpi, as determined using ELISA. N = 5. Mean ± SEM. No significant difference between viable and avitalized treatments. Asterisks *, ** and *** denote significance levels of p < 0.05, p < 0.01, or p < 0.001, resp.

Figure 7. Co-injection of recombinant TNF-α does not improve bacterial anti-tumor effects in more rigid tumor models

Tumor-bearing mice were inoculated i.v. with 5 × 10⁶ Salmonella SL7207 or E. coli probiotic Symbioflor-2 in concert with 1 μg recombinant TNF-α. Therapeutic efficacies were compared. (A) Levels of TNF-α in sera at 1.5 hour post treatment, as determined using ELISA. N = 5. Combination therapy yields excessive levels of TNF-α. (B) RenCa tumor development profiles in treated WT mice. N = 8. (C) F1.A11 tumor development in treated WT mice. N = 5. (D) Endpoint RenCa tumor volume. Comparison at 9 dpi. (E) Endpoint F1.A11 tumor volume. Comparison at 10 dpi. TNF-α supplementation does not improve bacterial therapeutic effects. Displayed are Means ± SEM. Asterisks *, ** and *** denote significance levels of p < 0.05, p < 0.01, or p < 0.001, resp.

Figure 8. While the infectious dose of Salmonella decides about effectiveness of colonization, TNF-α supplementation supports tumor invasion at lower inoculi

CT26 tumor-bearing mice were inoculated i.v. with Salmonella SL7207 alone or in combination with 1 μg recombinant TNF-α. Infectious doses were titrated from 5 ×10⁶, decreasing in increments of 1 log down to 5 ×10³. (A) Level of active TNF-α in sera as determined using a TNF-α sensitive C5F6 fibroblast bioassay. N = 5. Mean ± SEM. Lower doses fail to induce a significant TNF-α response. Exogenous rTNF-α can yield compensatory levels. (B) Frequency of colonized CT26 tumors. Results were pooled from three separate experiments. (C) CT26 tumor colonization 48 hpi. Tumors not colonized (refer to Figure 8B) were excluded from this calculation. Displayed are Medians ± range. Lower doses substantially reduce the frequency and efficiency of tumor colonization, compensable through co-injection of rTNF-α. (D) Splenic colonization. N = 5. Median ± range. Gradual decrease in adverse colonization with lower infectious doses and no marked change upon supply of rTNF-α. Asterisks *, ** and *** denote significance levels of p < 0.05, p < 0.01, or p < 0.001, resp.