1. Abstract
Among the six known functional human beta-tubulin genes, the TUBB gene is widely expressed and thought to encode the major tubulin in epithelial tumor cells with which taxanes interact. Mutation of TUBB has been reported in Chinese hamster ovary cells and an ovarian tumor cell line adapted for growth in the presence of paclitaxel. In NSCLC, Monzó et al. have reported TUBB mutation in 16 of 49 (33%) tumor samples from untreated patients and found a striking association between the presence of mutation and both poor treatment response to paclitaxel and shortened overall survival (JCO 17:1786). Based on this observation, the presence of beta-tubulin mutation has been proposed as the basis for chemotherapeutic drug selection in the management of patients with advanced NSCLC. To better study the association of TUBB mutation with tumor cell growth and taxane resistance, we analyzed TUBB in 25 NSCLC cell lines. PCR amplification of the four coding exons of TUBB from genomic DNA was accomplished by selecting oligonucleotide primers for PCR such that at least one oligonucleotide was complementary to intronic sequence. This requirement avoids co-amplification of the several tubulin processed pseudogenes that have high sequence identity within the coding region but are devoid of introns. Purified PCR products were sequenced on one strand for exons 1, 2, and 3 (57, 109, and 101 bases, respectively) and on both strands for exon 4 (1068 bases). Only two sequence variants were identified in the coding region and adjacent splice sites. Both were heterozygous single nucleotide variants in the third position of codons 187 and 217 that did not change the predicted amino acid. Each variant was present in one cell line and likely represents a polymorphism. Thus, mutation of TUBB is not common in NSCLC cell lines. Analysis of tumor samples is ongoing, however, the presence of TUBB mutation in tumor samples but not cell lines would be unusual in NSCLC where the frequency of genetic alterations of other genes is generally greater in cell lines than tumor samples. Alternatively, the previously reported TUBB mutations in NSCLC may be an artifact of co-amplificaiton of pseudogenes.

2. Introduction (1)
Tubulin, the cellular target for taxanes, is composed of ab heterodimers.
There are at least 6 genes encoding beta subunits of tubulin.1 In most epithelial tumor cells, TUBB is the most highly expressed isoform.
Chinese hamster ovary cells and an ovarian tumor cell line adapted for growth in vitro in the presence of paclitaxel have been found to have mutations of TUBB4,5

3. Introduction (2)
Monzó et al.2 have reported TUBB mutation in 16 of 49 (33%) tumor samples from untreated patients with NSCLC (all but two in exon 4).
They also found a significant association between the presence of mutation and both poor treatment response to paclitaxel and shortened overall survival.2
Presence of TUBB mutation has been proposed as a basis for selecting initial chemotherapy for patients with advanced NSCLC

4. Methods (1) NSCLC cell lines
Twenty-five NSCLC tumor cell lines were selected for tubulin gene analysis on the basis of available in vitro taxane drug sensitivity. Genomic DNA was isolated from these cell lines by proteinase K digestion and phenol/chloroform extraction.

5. Methods (2) Oligonucleotides for PCR
Oligonucleotides were designed based on genomic sequence from the TUBB region from the Human Genome Project (GenBank accession number AC006165) using GeneWorks version 2.45 and requiring at least one oligonucleotide to be within intronic sequence.3

6. Methods (3) 1F CCCATACATACCTTGAGGCG
1R TTTGGACCGTTAGAAGCCC (sequencing oligo)
2F2 GAAGCAGAGGTTGCAGTGAG
2R2 TGACAGATTCACCCAAAGGG
2F AGAGCGAGACTCCGTCTCAA (sequencing oligo)
3F TCCCTTCTGCCAGATTTCAC
3R2 CAGGACAGAATCAACCAGCTC
3R CCCCTACTGCCCCATAATTT (sequencing oligo)
4F3 AGGTAGTGCCTACTATTGCTGG
4R4 TGAGTAAGACGGCTAAGGGAAC (sequencing oligo)
4R2 AGCCATCATGTTCTTGGCA (sequencing oligo)
4F2 AGTTGGCAGTCAACATGGTC (sequencing oligo)
4F4 TTGAGCTTTTCTCCTGACTGC (sequencing oligo)
TP4F AGAGAGCTGTGACTGCCTG (from Monzó, 1999)
TP4R AAGGTATTCATGATGCGA
STP4 TCAGGGTATTCTTCTCGGAT (sequencing oligo)

7. Methods (4) TUBB amplification and sequencing3
PCR was performed in 10 mL reaction volume containing 1X PCR buffer, 1.5 mM MgCl2, 5 units Taq platinum, 0.2 mM each dNTP, and 0.2 mM each oligonucleotide primer. Typical cycling conditions were initial denaturation of 94C for 2 min followed by cycles of 94C for 30 sec, annealing temperature for 30 sec, and 72C for 2 min and final extension time of 10 min at 72°C. PCR products were purified using QIAquick Spin PCR Purification Columns per manufacturer’s recommendations. DNA was sequenced using BigDye terminator kit.

8. Results (1) Summary of sequencing of TUBB gene
exons 1 to 4 in 25 NSCLC cell lines
Cell lines with variants
Wild type cell lines (amino acid change)
H23 H1385 H1930 H1648 (L187L*)
H322 H1437 H1944 H2228 (L217L*)
H522 H1466 H2023
A549 H1650 H2122
H676 H1755 H2286 * both silent changes
H835 H1975 H2342
H1264 H1793 H2409
H1373 H1869

9. Results (2) Sequence variant in cell line H2228 using intron primers
Sense
Antisense

10. Results (3) Why do these results differ from those of Monzo?
Comparison of sequence of PCR products using intronic primers with those using exonic primers

11. Results (4) Genomic Organization of TUBB Exon: 1 2 3 4
Oligos from this work:
4R2
4R4
4F3
4F4
4F2
Exon 4
intron 3
3’ UT
TB4-F
Oligos from Monzo:
TP4-F
TR4-F
STR4
STB4
TB4-R
(location unclear)
TP4-R
TR4-R
STP4

12. Results (5) Location of 5’ PCR primer Intron Exon
Codons
Sequence variants are observed when primers are located in Exon only

13. Results (6)
Sequence variants found in the ribose-binding region of TUBB from
tumor cell line DNA amplified with PCR primers located within exons.
bp change codon codon change AA change type of change
474 A to A/G GAA-GAG Glu-Glu silent
484 C to C/T CGC-TGC Arg-Cys missense
487 A to A/T ATC-TTC Ile-Phe missense
510 G to G/T GTG-GTA Val-Val silent
518 C to C/A CCC-CAC Pro-His missense
522 A to A/G AAA-AAG Lys-Lys silent
538 G to G/A GTC-ATC Val-Ile missense
540 C to C/T GTC-GTT Val-Val silent
544 C to C/T CCC-TCC Pro-Ser missense
555 C to C/T GCC-GCT Ala-Ala silent
585 T to C AAT-AAC Asn-Asn silent
603 C to C/T TGC-TGT Cys-Cys silent
612 C to C/T AAC-AAT Asn-Asn silent

14. TUBB Pseudogenes (1)
A pseudogene is a sequence which is homologous to functional genes but which contains mutational changes precluding the formation of a functional product.6,7
Processed pseudogenes have structure similar to mRNA: no introns, no promoter, and with poly(A) tract.6,7
Genomic Organization of a TUBB pseudogene
AATAAAN14(A)n
11-bp direct repeat

15. TUBB Pseudogenes (2)
There are at least 8 pseudogenes for TUBB in the human genome (see alignment of ribose binding domain at right). At least some of these sequences are co-amplified when using primers within exons.
Alignment of the ribose binding domain of exon 4 of TUBB with 8 sequences from the human genome that may be co-amplified with TUBB using exonic primers is shown at right.

16. Conclusions (1) Mutation of TUBB is uncommon in NSCLC.
Prior reports of TUBB mutation in NSCLC may result from simultaneous amplification of TUBB, related tubulin genes, and processed pseudogenes.

17. Conclusions (2)
TUBB mutation should not be used as a basis for selection of chemotherapy in NSCLC.
The apparent association of sequence variants detected by TUBB sequence analysis with tumor response and survival in NSCLC remains unexplained.

18. References 1. Ludueña, RF Int Rev Cytolol 178:207, 1998.
2. Monzó, M et al. JCO 17:1786, 1999.
3. Owshalimpur D et al. Molec Probes, 1999.
4. Giannakakou, P, et al, JBC 272:17118, 1997.
5. Gonzalez-Garay, ML, JBC 274:23875, 1999.
6. Wilde, CD et al. Nature 297:83, 1982.
7. Gwo-Shu, M et al. Cell 33:477, 1983.