Mechanisms of Resistance to Gemcitabine: Comparison
View Latest Version
Please note this is a comparison between Version 1 by Stephan Reshkin and Version 4 by Sirius Huang.

Gemcitabine is still the standard-of-care chemotherapeutic drug for pancreatic ductal adenocarcinoma (PDAC). However, the response rate is quite low. There are multiple mechanisms and participants in gemcitabine resistance.

  • resistance to treatment
  • gemcitabine
  • desmoplastic reaction
  • hydroxyurea
  • proteasome inhibitors

1. Gemcitabine

Gemcitabine was introduced in pancreatic cancer treatment in 1997 after Burris et al. published their report [1][214]. This was a randomized clinical trial of 126 patients with advanced pancreatic cancer. They found that gemcitabine achieved better results than fluorouracil (5-FU) regarding a modest overall survival improvement and pain control. The mean survival only improved by one month (5.65 for gemcitabine vs. 4.41 for 5FU), but improvements in pain and the Karnofsky index were significant (23.8% for gemcitabine vs. 4.8% for 5FU).
Gemcitabine is still the standard-of-care chemotherapeutic drug for pancreatic ductal adenocarcinoma (PDAC) [2][3][215,216]. However, the response rate is quite low (around 30%), and even lower in advanced cases [4][5][217,218].
It improves average survival by two to three months [6][219], a really poor result. Chemoresistance develops rapidly [7][220] and is therefore the main limiting factor of the drug.
Gemcitabine is used as monotherapy or in combination with other chemotherapeutic drugs [8][221]. Results in combinatorial treatments are slightly better than monotherapy, however the high toxicity involved in combinatorial schemes led to its being used alone in many cases.

2. Mechanisms of Resistance to Gemcitabine

There are multiple mechanisms and participants in gemcitabine resistance. The following factors have been identified:
1. Decrease in deoxycytidine kinase expression or activity [9], thus impeding gemcitabine activation in this rate limiting step [10][11];
  1. Decrease in deoxycytidine kinase expression or activity [264], thus impeding gemcitabine activation in this rate limiting step (Figure 4) [265,266];
  2. Increased expression of Ribonucleotide Reductase isoform M1 [267,268,269], leading to increased production of nucleotides for DNA synthesis;
  3. Activation of the PI3K/Akt survival pathway with its anti-apoptotic effect [270];
  4. Downregulation of the hypoxia-induced pro-apoptotic gene BNIP3 [271];
  5. Over-expression of focal adhesion kinase (FAK) [272];
  6. Over-expression of Src kinase [273,274];
  7. c-Met activation: inhibition of c-Met with cabozantinib has overcome gemcitabine resistance and increased its cytotoxicity [275,276,277,278,279];
  8. Over-expression of the transcription enhancer high-mobility group A1(HMGA1) [280,281,282];
  9. Deoxycytidine release from stellate cells [283];
  10. CAF-released exosomal miRNA 106-b [284]: CAFs are intrinsically resistant to gemcitabine and transmit this resistance through exosomes containing miRNA 106-b to cancer cells where they target TP53; For a review on miRNAs in pancreatic cancer, read Slotowinski et al. [285].
  11. CAF production ofthe chemokine stromal cell-derived factor 1 (SDF1) is able to activate special AT-rich sequence-binding protein 1 (SATBP1), which intervenes in tumor progression and resistance to gemcitabine [286], as shown in the lower panel of Figure 8;
  12. SDF-1α produced by stellate cells and secreted in the stroma has the ability to bind the CXCR4 over-expressed in pancreatic cancer cells, activating a pathway that increases survival, reduces apoptosis, and increases expression of IL-6 [287], as shown in Figure 8.
  13. TRIM 31 expression by activating NF-kB [288];
  14. TGFβ1 also induces gemcitabine resistance [289];
  15. Epithelial–mesenchymal transition [290], as the relationship between EMT and gemcitabine resistance is very complex:
    • Gemcitabine-resistant cells acquire EMT phenotype with cancer stem cell characteristics. Notch-2 and Jagged-1 are highly upregulated in these cells [291];
    • Gemcitabine resistance-mediated EMT is in part induced by hypoxia because when HIF-1α is blocked, there is partial reversal of EMT [292];
    • miR 233 is a contributing factor to gemcitabine resistance-dependent EMT [293];
    • Cells that survive after gemcitabine treatment show increased stemness and EMT markers [294];
    • Gemcitabine-induced EMT sustains chemoresistance [295];
    • Targeting EMTcan overcome resistance [
; 10. CAF-released exosomal miRNA 106-b [29]: CAFs are intrinsically resistant to gemcitabine and transmit this resistance through exosomes containing miRNA 106-b to cancer cells where they target TP53; For a review on miRNAs in pancreatic cancer, read Slotowinski et al. [30]. 11. CAF production of the chemokine stromal cell-derived factor 1 (SDF1) is able to activate special AT-rich sequence-binding protein 1 (SATBP1), which intervenes in tumor progression and resistance to gemcitabine [31], as shown in the lower panel of Figure 1; 12. SDF-1α produced by stellate cells and secreted in the stroma has the ability to bind the CXCR4 over-expressed in pancreatic cancer cells, activating a pathway that increases survival, reduces apoptosis, and increases expression of IL-6 [32], as shown in Figure 1. 13. TRIM 31 expression by activating NF-kB [33]; 14. TGFβ1 also induces gemcitabine resistance [34]; 15. Epithelial–mesenchymal transition [35], as the relationship between EMT and gemcitabine resistance is very complex:
  • Gemcitabine-resistant cells acquire EMT phenotype with cancer stem cell characteristics. Notch-2 and Jagged-1 are highly upregulated in these cells [36];
  • Gemcitabine resistance-mediated EMT is in part induced by hypoxia because when HIF-1α is blocked, there is partial reversal of EMT [37];
  • miR 233 is a contributing factor to gemcitabine resistance-dependent EMT [38];
  • Cells that survive after gemcitabine treatment show increased stemness and EMT markers [39];
  • Gemcitabine-induced EMT sustains chemoresistance [40];
  • Targeting EMTcan overcome resistance [41].
    • ].
2. Increased expression of Ribonucleotide Reductase isoform M1 [12][13][14], leading to increased production of nucleotides for DNA synthesis; 3. Activation of the PI3K/Akt survival pathway with its anti-apoptotic effect [15]; 4. Downregulation of the hypoxia-induced pro-apoptotic gene BNIP3 [16]; 5. Over-expression of focal adhesion kinase (FAK) [17]; 6. Over-expression of Src kinase [18][19]; 7. c-Met activation: inhibition of c-Met with cabozantinib has overcome gemcitabine resistance and increased its cytotoxicity [20][21][22][23][24]; 8. Over-expression of the transcription enhancer high-mobility group A1(HMGA1) [25][26][27]; 9. Deoxycytidine release from stellate cells [28]
Cancers 14 02486 g008 550
Figure 18.  The two pathways shown in the figure have been found to decrease gemcitabine’s cytotoxicity and apoptosis. SDF-1α expression is induced by galectin 1.
Therefore, based on the above evidence, a circuit like the one shown below may represent the chain of events:
Gemcitabine→ EMT→ Resistance to Gemcitabine→ further EMT
 
16.
ATP binding cassette (ABC) re-exports cytotoxic compounds in general, including gemcitabine [42][43][44][297,298,299];
17.
CXCL12-CXCR4 signaling [45][46][47][300,301,302];
18.
miRNA 320c through SMARCC1 (SMARCC1 is a protein that forms part of the SWI/SNF complex) [48][303]: This miRNA exerts contradictory actions because it has anti-tumor effects in bladder cancer by downregulating CDK6 [49][304] and in glioma, where it decreases growth and metastasis [50][305]. It was found to decrease canonical Wnt signaling in joints [51][306]. Therefore, miRNAwe 320c can becan considered miRNA 320can anti-oncogenic miRNA [52][307] which, however, promotes gemcitabine resistance.
19.
miRNA 21 and 221 [53][54][308,309]: miRNA 21 binds to the 3′-UTR region of the Bcl-2 gene, leading to its over-expression and thus inhibiting apoptosis of pancreatic cancer cells [55][310]; antisense miRNAs 21 and 221 restored gemcitabine sensitivity and induced cell-cycle arrest and apoptosis [53][308], while miRNA 200 [56][57][311,312] seems to antagonize miRNA21. miRNA 221 is considered a reliable circulating miRNA for diagnostic purposes [58][313];
20.
miRNA 155 modulates exosome synthesis and promotes gemcitabine resistance [59][314]; prolonged treatment with gemcitabine increased miRNA 155 levels, which in turn increased exosomes and expression of anti-apoptotic proteins. The message is carried to the rest of the cells through exosomes.
21.
miRNA 99a and miRNA100 [60][315] have been proposed as prognostic markers of gemcitabine resistance;
22.
miRNA 214 [61][62][316,317]: Low expression of miRNA 214 was predictive for improved results of the gemcitabine–vinorelbine association in metastatic esophageal cancer [63][318];
23.
miRNA 365 induces gemcitabine resistance by targeting the anti-apoptotic BAX protein and its adaptor protein SHC1 [64][319]. It also induces the production of survival-related proteins;
24.
miRNA 210 [65][320] downregulates Homeobox protein Hox-A9, which increases NF-kB activity and decreases sensitivity to gemcitabine. However, the role of this microRNA is controversial. Amponsah et al. [66][321] identified miR-210 as a direct suppressor of the multidrug efflux transporter ABCC5; miR210 probably has dualpro- and anti-tumoral effects according to the balance with the oncomucin MUC4. They mutually regulate each other [67][322];
25.
miRNA-17-5p is usually overexpressed in PDAC, participating in carcinogenesis and tumor progression [68][323] and inhibiting Bim expression, thus decreasing apoptosis. Experimental inhibition of miR-17-5 increased sensitivity to gemcitabine [69][324];
The effects of some miRNAs regarding gemcitabine resistance are still a matter of debate. This is the case of miR-421, which seems to be pro-tumoral, reducing the expression of DPC4/SMAD4 [70][325] and at the same time increasing gemcitabine efficacy through decreased SPINK1 expression [71][326]. Furthermore, there is an oleic acid derivative, K73-3, that is able to upregulate miRNA 421 in vitro and in vivo, improving gemcitabine cytotoxicity [72][327]. Therefore, miRNA 421 should be considered an anti-oncogene agent.
In addition to the miRNAs discussed above, there are others without fully proven inhibitor effects on gemcitabine:
26.
MUC1 and MUC4 [73][74][328,329] (Figure 29): Oncomucins play an important role in gemcitabine resistance that is discussed below. Mucins form a protective envelope surrounding cancer cells and participate in chemoresistance by impeding drug access to the malignant cells. Their production is usually highly increased in pancreatic cancer. There are two mechanisms involved in oncomucin-induced gemcitabine resistance:
Cancers 14 02486 g009 550
Figure 29. The two mechanisms involved in treatment resistance induced by oncomucins. This diagram is based on references [75][76][77][78][79][80][81][82][83][84][85][86][87][88][330,331,332,333,334,335,336,337,338,339,340,341,342,343].
(1)
 Direct, by MUC1 inhibiting the apoptotic BAX protein and increasing stemness;
(2)
 Indirect, by inducing Her 2 signaling.
Tumor-associated oncomucins have a different glycosylation pattern. MUC1 is less glycosylated than MUC4. MUC1-C, the intracellular portion of MUC1, is a driver for the upregulation of PD-L1. Although this immunoescape was found in triple-negative breast cancer, itwe can be hypothesized that pancreatic cancer’s refractoriness to immune-checkpoint inhibitors may be related to MUC1-C. MUC1-C expression also protects the malignant cells against genotoxic attacks in general;
MUC5AC, a facilitator of migration and invasion, also participates in drug resistance by inhibiting TRAIL death pathways.
27.
According to Shukla et al. [89][344], HIF-1α-dependent high glycolytic flux is the main player in gemcitabine resistance. High glycolytic flux allows for a high cytidine pool that competes with gemcitabine;
28.
CD44-expressing cells are resistant to gemcitabine. MDR1 is overexpressed in these cells [90][345]. These CSCs can rebuild the tumor after chemotherapy;
29.
Tumor heterogeneity: gemcitabine was more effective on cells that were more than 400–500 mμ from the desmoplastic areas [91][346]. Interestingly, high doses of metformin killed cells closer to the desmoplastic reaction area;
30.
ROCK2 (Rho associated protein kinase 2) activity is a cause of acquired gemcitabine resistance [92][347]. The Rhoa/ROCK2 axis promotes migration and metastasis. A pathway has been found in PDAC that shows the long non-coding RNA ZFAS1 inducing metastasis through the Rhoa/ROCK2 axis [93][348]. ZFAS1 is usually overexpressed in PDAC.ROCK inhibitors sensitize pancreatic CSCs to gemcitabine [94][349] and also reduce metastasis;
31.
Constitutive activation of NF-kB [95][96][350,351]: IL-1α expression is induced by NF-κB, which in turn increases NF-kB in a positive feedback loop, leading to permanent NF-kB activity [97][98][99][100][352,353,354,355]. In addition to the classic PI3K/AKT/NF-kB pathway that is fully operative in PDAC, several other pathways that induce gemcitabine resistance through NF-kB activity have been identified [101][356];
Pancreatic tumors show low miRNA 146-5p expression, impeding regulation of the TRAF6′s 3 UTR segment, thus allowing the pathway shown above. (TRAF6 is the tumor necrosis factor receptor-associated factor 6 that works as an adaptor protein allowing protein–protein interactions) [102][357].
PARP 14 (Poly ADP-ribose polymerase) is highly expressed in PDAC and is associated with poor prognosis. Silencing PARP 14 reduced resistance to gemcitabine.
Clusterin is a protein associated with chemoresistance to different chemotherapeutics. It was found to be increased in PDAC [103][358].
In summary, independently of which pathway activates NF-kB, this transcription factor has the ability to eliminate the pro-apoptotic effects of gemcitabine. Blocking NF-kB can, to a certain degree, decrease gemcitabine resistance [104][359].
32.
Increased expression of heme-oxygenase-1 (HO-1): PDAC cells show a 6-fold expression of HO-1 compared with normal pancreatic cells. Gemcitabine and/or radiotherapy treatments further increases HO-1 expression. HO-1 knockdown increases sensitivity to both therapies [105][360];
33.
Decreased expression of hENT1, the gemcitabine transporter, reduces its intracellular access [106][361];
34.
High expression of the polo-like kinase [107][362]: Downregulation of this kinase decreases resistance [107][108][362,363]. Rigosertib, a multikinase inhibitor, has been developed for this purpose. It is being tested in clinical trials [109][364];
35.
Decreased glutathione peroxidase 1 induced resistance to gemcitabine: Glutathione peroxidase 1 modulates the AKT/GSK3β/Snail signaling axis in PDAC [110][365]. Interestingly, gemcitabine is able to induce the expression of glutathione pathway-related genes which are suspected of generating resistance [111][366];
36.
Increased expression of Snail [112][367];
37.
Increased survivin expression [113][368]: Emodin, an inhibitor of survivin expression, increases gemcitabine cytotoxicity [114][369] and similar results can be obtained with small interference RNA (siRNA) [113][368];
38.
Decreased intracellular ceramide/sphingosine-1-phosphate [115][370]: Increased ceramide favors apoptosis, while increased sphingosine-1-phosphate is anti-apoptotic; sphingosine kinase-1 is the enzyme that controls this ratio generating sphingosine-1-kinase, thus exerting anti-apoptotic effects;
39.
Mutation or deletion of the BRCA2 gene [116][371];
40.
Activation of Notch signaling [36][117][118][291,372,373] increases therapeutic resistance: This is related to the acquisition of an epithelial-mesenchymal phenotype (see paragraph 15); downregulation of Notch signaling has a chemosensitizing effect [119][374]; Notch-induced chemoresistance to gemcitabine is partly the result of Notch’s ability to alter the intrinsic apoptotic pathway [120][375];
41.
Hedgehog signaling [121][376]: chemotherapy activates the Hedgehog pathway [24][279], and this activation in turn leads to the expression of stem cell markers such as CD44, SOX2, OCT4, Nanog, and drug efflux proteins of the ATP-binding cassette family. Thus, Hedgehog increases stemness and induces a multidrug resistance phenotype [122][377];
42.
Cytosolic 5′-nucleotidase 1A over-expression [123][378]: This enzyme is able to reduce gemcitabine’s intracellular metabolites [124][379]. The histone deacetylase inhibitor trichostatin A has been found to synergize with gemcitabine, increasing its cytotoxicity, and importantly, inhibiting 5′-nucleotidase [125][380];
43.
Pancreatic cancer stem cells [126][381]: Stemness is a key factor in therapeutic failure in most tumors. CSCs do not respond to chemotherapy and are able to replicate the tumor after cytotoxic destruction of sensitive cells. The activation of pancreatic cancer stem cells has shown abilities to promote resistance to gemcitabine. Many of the activators are also involved in resistance. (Figure 310);
Cancers 14 02486 g010 550
Figure 310. The yellow squares are the known activators of pancreatic cancer stem cells. The other activators (white squares) have also been found to play a role. This diagram is based on references [127][128][129][130][131][132][133][134][135][136][137][138][139][140][141][142][143][144][145][146][382,383,384,385,386,387,388,389,390,391,392,393,394,395,396,397,398,399,400,401]. Importantly, many of the stemness activators are also involved in epithelial–mesenchymal transition.
44.
Pancreatic cancer stromal stem cells [147][148][402,403];
45.
Calcyclin-binding protein or Siah-1-interacting protein (CacyBP/SIP) was found to be overexpressed in MDR after gemcitabine treatment. This protein induced P-gp and BCL2 expression reducing apoptosis [149][404]. In addition, CacyBP/SIP knockdown suppresses proliferation in pancreatic cancer by downregulating cyclin E and CDK2 and upregulating Rb and p27 [150][405];
46.
Soluble V-CAM, produced by pancreatic cancer cells, recruits tumor-associated macrophages (TAMs) [151][406];
47.
De novo lipid synthesis [152][407];
48.
The extracellular matrix composition: Laminin and collagen type IV-ECM (mimicking an early tumor ECM) protects from drug-induced apoptosis compared to a collagen I-rich late-tumor ECM;
49.
Increased galectin 1 expression in stellate cells [153][154][155][156][408,409,410,411]: MiRNA 22 was found to reduce the expression of galectin1 in hepatocellular carcinoma [157][412];
50.
Autophagy has been shown to be upregulated in PDAC and it plays an important role in resistance to chemotherapy [158][159][413,414]. Autophagy is an inducer of gemcitabine resistance and is probably one of the mechanisms that cells use to survive cytotoxic drugs. Gemcitabine’s cytotoxicity is increased when an autophagy-inhibitor is used simultaneously [160][415]. Pancreatic adenocarcinoma is a very hypoxic tumor, and hypoxia can induce autophagy. Additionally, the expression of high-mobility group box 1 (HMGB1) is an autophagy inducer. Interestingly, gemcitabine upregulates this protein, thus indirectly increasing autophagy [161][416]. In a preoperative setting, when combining the autophagy inhibitor hydroxychloroquine with gemcitabine, 61% of patients showed CA19.9 marker decrease, improved postoperative, and disease-free survival. These findings were particularly evident in the patients with high levels of the autophagy marker LC3-II [162][417]. By blocking autophagy, gemcitabine’s cytotoxic effects were increased and stem cell activity reduced [163][418]. Zeh et al. [164][419] studied two cohorts of preoperative patients, one receiving nab-placlitaxel, and another group with the same medication plus hydroxychloroquine. They found that the resected pancreas in the hydroxychloroquine group had a greater pathologic response and higher immune activity. However, overall survival and disease-free survival was similar in both cohorts. SNHG14 (small nuclear RNA host gene 14) oncogene expression generates a long non-coding RNA that induces autophagy and resistance to gemcitabine [165][420]. This LNC-RNA seems to act as an anti-sense against MiRs involved in anti-tumoral activity;
The conclusion is that there is evidence supporting better results with longer overall survival and disease-free survival by adding autophagy inhibitors to gemcitabine in the resectable cases [166][421]. Evidence in this regard is lacking for inoperable patients.
51.
Pancreatic tumor microbiota: There is a clinically important population of bacteria and fungi within the pancreas and biliary tree in patients with PDAC and this population is different from the microbiota found in the normal pancreas [167][168][169][422,423,424]. The bacteria present in PDAC show some specificity [170][425]. Regarding gemcitabine, it was found that intratumoral Gammaproteobacteriahad a role in resistance [171][426]. Patients that had some surgical or endoscopic procedure on the pancreas and the biliary tree were prone to host pro-resistance bacteria in the pancreas [172][427], and 5-FU resistance was associated with the presence of Fusobacterium nucleatum in colorectal cancer [173][428]. Fusobacterium is very abundant in PDAC, so it can be hypothesized that it also plays a role in pancreatic chemoresistance. Furthermore, Fusobacterium induces autophagy as part of its chemoresistance mechanism, another frequent finding in PDAC. Fungi have also been found to be a possible cause of gemcitabine resistance [174][429];
52.
Hypoxia is a key factor in the PDAC phenotype, including proliferation, autophagy, progression, metastasis, as well as resistance to treatment in general, and to gemcitabine in particular [175][430]. The evidence is compelling [176][177][178][179][180][181][182][183][431,432,433,434,435,436,437,438]. A simple example shows the importance of this issue. Hypoxia is expressed through the hypoxia-inducible factors (transcription factors that modulate over 150 genes). Downregulation of HIF-1α with a newly developed molecule, LW6, inhibited autophagic flux, improved the efficacy of gemcitabine, stopped proliferation, and induced cell death [184][439]. LW6 is a novel HIF-1 inhibitor that decreases HIF-1α protein expression [185][186][187][440,441,442];
Hypoxia not only increases resistance to gemcitabine, it also increases gemcitabine-induced stemness [188][443]. Luo et al. [189][444] showed that hypoxia- induced miRNA 301a which in turn promoted gemcitabine resistance through downregulation of T53, thereby integrating hypoxia, miR, and gemcitabine resistance into one pathway. Figure 11.
Cancers 14 02486 g011 550
Figure 411.
Mechanism of hypoxia-induced gemcitabine resistance.
53.
Increased expression of cytoplasmic ribonucleotide reductase subunit M1 (RRM1) [190][445]: RR is a multimeric enzyme essential for maintaining a high pool of deoxynucleotides for DNA elongation and also for DNA repair. Gemcitabine-resistant pancreatic cancer cells treated with RRM1 inhibitors showed considerable decrease in resistance [13][268]. Patients with high RRM1 levels showed a poorer overall survival with gemcitabine treatment compared with low RRM1- expressing patients [191][446];
54.
The Hippo pathway is involved in organ size control and tissue homeostasis. It was found that this pathway plays a role in drug resistance [192][193][115,447].
Figure 512 presents a summary of mechanisms involved in gemcitabine resistance.
Cancers 14 02486 g012 550
Figure 512. Some mechanisms of resistance to gemcitabine in PDAC. ABC: ATP binding cassette. Poor vascular supply and the desmoplastic reactionare mainly physical barriers. The numbers are chemicals and pathways activated for the escape. ABC re-exports the cytotoxic substances.
© Text is available under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/)
Video Production Service
Feedback